SAN FRANCISCO, May 28 — Meta yesterday launched paid subscription plans for its flagship apps, marking a major push by the tech giant to diversify beyond its longtime reliance on advertising revenue.

Meta head of product Naomi Gleit announced the move in a video posted to Instagram, saying she was rolling out Facebook Plus, Instagram Plus and WhatsApp Plus globally, with more plans in the works for businesses, creators and artificial intelligence products.

The move comes as Meta faces investor scrutiny over its massive AI spending. The company has projected capital expenditure—mainly for AI data centers—of between US$125 billion and US$145 billion for the year.

Meta’s stock rose nearly three percent on the news.

According to reports, Instagram Plus and Facebook Plus will be priced at US$3.99 per month, while WhatsApp Plus will cost US$2.99 per month.

Instagram Plus and Facebook Plus will unlock extra features including better analytics, story rewatch statistics, wider audience reach and profile customization options.

WhatsApp Plus focuses on personalization, including premium stickers, custom ringtones and app themes.

Gleit said the company intends to eventually consolidate its various offerings under a single brand called Meta One.

Meta in 2023 launched ad-free, paid versions of Facebook and Instagram in Europe to comply with EU data privacy legislation, giving users a choice between a free, ad-supported experience and a paid, ad-free one. — AFP

DHAKA, May 25 — Bangladesh’s densely populated capital, infamous for grinding gridlock, has launched its first AI-powered traffic enforcement system in a bid to bring order to one of the world’s most congested cities.

Across Dhaka, buses, cars, motorcycles and pedal rickshaws jostle for space — seeming to treat signals, pedestrian crossings and even traffic police as obstacles to overcome.

On-the-spot police fines often spark arguments, confrontations are common, and officers are sometimes struck by vehicles refusing to obey commands.

“Those who violate the rules turn against us,” traffic sergeant SM Nazim Uddin told AFP.

“But since AI was introduced, people behind the wheel have started obeying the law — and we have been spared the everyday quarrels.”

Despite repeated attempts to modernise traffic management, the city of more than 22 million people had remained reliant on manual control, with officers stretching ropes across roads before traffic lights turned green.

But in April, Dhaka police linked traffic cameras to artificial intelligence software designed to automatically detect violations.

Motorist Hannan Rahman Jibon, 28, was among the first drivers caught.

“I ran a red light and the owner of my car, who was sitting at home, received a text message saying that the vehicle had violated traffic rules,” said Jibon, who was hit with an automatic 2,000 taka (RM65) fine.

“I am more careful now, with cameras installed in many different parts.”

People travel past an AI-based traffic management system (C) installed along a road in Dhaka May 14, 2026. — AFP pic

‘Getting results’ 

One study, by the US National Bureau of Economic Research, ranked Dhaka as the “slowest city” in the world.

With an average speed of just 4.8 kilometres per hour, according to a World Bank and Bangladesh University of Engineering and Technology study, it is quicker to walk.

City police spokesman NM Nasiruddin said the AI system used existing traffic-monitoring camera feeds — with the software identifying offences ranging from signal and lane violations to illegal parking.

Coverage does not stretch across all the city, and manual traffic enforcement will be phased out over time, but police say it is changing behaviour.

“We have started getting results,” Nasiruddin said. “We have prosecuted at least 300 vehicles.”

Inside a control room at police headquarters, analyst Sharmin Afroze, 52, monitored a bank of live traffic feeds, plugged into the AI system, where violations it flags are checked by humans.

“Before, police used to stop vehicles, check papers and determine fines,” Afroze told AFP, noting that in a single day, the system had recorded nearly 800 traffic violations.

For now, police are only fining the worst offenders — while giving the others a warning.

City police spokesman NM Nasiruddin poses for a photograph at his office in Dhaka May 14, 2026. — AFP pic

‘Do not follow rules’ 

Still, officials admit the technology faces hurdles.

“We are facing some challenges too,” Afroze said. “For example, some number plates are blurred or too small to identify.”

Police are working with the road transport authority to address the issue, while additional features — such as vehicles driving on footpaths — will be added soon.

The system also currently targets motor vehicles, and police are still considering how to regulate Dhaka’s vast fleet of pedal rickshaws, a defining feature of the city’s traffic.

Hasib Mohammed Ahsan, a professor at Bangladesh University of Engineering and Technology, said the long-term success of the system would depend less on technology — and more on whether authorities consistently enforce the rules.

“We have spent huge amounts of money on traffic signals and their upgrades, but those efforts were never sustained,” Ahsan said.

“We do not follow rules, there is no consistency in our planning, and there is no accountability for failure.” — AFP

PARIS, May 24 — With people increasingly adopting AI to help plan their vacations, hotels are working to make sure that you check them out — and check in.

Whether using ChatGPT or AI-enabled travel sites like Layla.ai, it is already possible to pose search questions like: “Calm hotel with west-facing balcony” or “Charming hotel with spa that accepts dogs”.

This simple switch to plain speech searches belies major technical changes that mean hotels have to learn to become visible to AI models.

“We’re in complete upheaval: last year 35 per cent of French people used artificial intelligence to find a hotel, a cafe or a restaurant,” said Nicolas Marette, founder of Custplace, a French company that helps firms optimise their digital presence.

According to a recent study by the Boston Consulting Group (BCG), around 37 per cent of travellers are already using AI-enabled online travel sites to plan and book trips.

Hospitality industry players have taken notice.

A quarter of hospitality firms “have an AI strategy that is starting to produce real returns across multiple organisational activities”, according to the BCG report.

“What a hotel needs to do to get well referenced by search engines is not the same thing that they need to do to get referenced by artificial intelligence,” said Johanna Benesty at BCG.

Moreover, not all AI models “work in the same way,” she added.

Plain speech, elaborate task 

At French hospitality group Accor, which owns dozens of chains including Pullman, Sofitel, Mercure and Ibis, “we’ve been trying for a year already to understand how to make ourselves more relevant... and be more visible,” the group’s AI and data science chief Nicolas Maynard told a recent industry conference.

But that can be a challenge as AI users see fewer options, meaning securing a top ranking becomes even more critical.

“It’s a big change: with Google a search gives you 50 results... while if you ask ChatGPT it will give you five” and that is it, Maynard added.

The switch to plain speech means big changes for hotels.

“The biggest challenge is to understand vague requests like ‘I want a romantic hotel in the south’,” Maynard said.

Because Accor’s systems do not currently classify properties by such attributes, the group has its work cut out.

“We need to adapt our systems to take semantics into account,” Maynard said.

Hyper detailed 

But beyond semantics, AI will allow hotels to provide customers with a wealth of information.

Best Western France’s director Olivier Cohn said he believed “what will make the difference is our ability to answer client questions more thoroughly”.

Hotels could respond to even the most detailed client questions such as “knowing if there is a power socket on the left side of the bed because they are used to sleeping on that side of the bed and charging their devices”, he said.

While such questions are simple in and of themselves, current systems and staff can struggle to answer in such detail, said Cohn, whose chain counts more than 4,000 hotels throughout the world.

Some hotels are already deploying AI chatbots to help answer simple guest questions, allowing staff to provide higher-value services.

But winning the referencing game isn’t only up to the hotels themselves.

BCG notes that “algorithms elevate properties with comprehensive, high-trust, multisource information over those with sparse or inconsistent digital footprints”, meaning that client descriptions and reviews will also be important.

But just like online travel agencies (OTA) charge commissions and offer premium service for a price, AI models are already beginning to do the same.

“The familiar OTA commission model will evolve into AI-era distribution fees, charged for prominence and relevance in algorithmic recommendations,” the BCG report said. — AFP

SOUTH PADRE ISLAND (United States), May 23 — SpaceX’s Starship spacecraft splashed down into the Indian Ocean Friday after the company performed a mostly successful test flight of the latest version of its enormous rocket.

The voyage was not without a few glitches, but SpaceX employees shown on a livestream roared in delight following the trial flight that comes as the firm owned by Elon Musk prepares a potentially record initial public offering.

The mammoth rocket blasted off into space at just after 5:30 pm local time (2230 GMT).

The company did not intend to recover the booster or the upper stage, and the final splashdown was fiery but controlled, as planned.

“Splashdown confirmed!” the company wrote on X.

SpaceX primarily aimed to demonstrate its redesigns in flight.

The third-generation Starship spacecraft carried out a maneuver that saw it flip upright and reignite its engines for control, despite one being out of commission.

It also deployed its 22 mock satellites, including two that attempted to photograph the spacecraft’s heat shield for analysis.

The vehicle had coasted through space but was not in exactly the correct orbit after one of its engines malfunctioned during an initial burn.

“I wouldn’t call it nominal orbital insertion,” company spokesperson Dan Huot said, adding however that it was “within bounds” of a previously analyzed trajectory.

After the Super Heavy booster separated from the upper stage as expected, Huot said on the livestream that the booster failed to complete its so-called boost-back burn.

The booster fell swiftly back to Earth, uncontrolled, into the Gulf of Mexico. SpaceX wasn’t planning to retrieve the booster anyway, but was still hoping for a precision return.

Musk applauded his team on X, calling the flight “epic.”

“You scored a goal for humanity,” he said.

‘Long way to go’

Friday’s flight followed an aborted trial one day prior.

The countdown clock stopped and started until it was determined that the last-minute red flags could not be addressed in time.

Musk quickly posted on X that “the hydraulic pin holding the tower arm in place did not retract.” 

SpaceX said that issue was corrected overnight.

The company is facing extra scrutiny after SpaceX filed earlier this week with US financial regulators to go public, likely in June, in what is expected to become a record IPO.

Friday marks Starship’s 12th flight overall, but the first in seven months.

The latest design is bigger than its predecessor, standing at just over 407 feet (124 meters) when fully stacked.

There’s a lot riding on SpaceX’s progress: the company is under contract with Nasa to produce a modified version of Starship to serve as a lunar landing system.

The US space agency’s Artemis program aims to return humans to the Moon, as China forges ahead with a rival effort that’s targeting 2030 for its first crewed mission.

Clayton Swope, an aerospace expert at the Center for Strategic and International Studies, told AFP that “the upgraded version of Starship did most of what SpaceX hoped it would do during the launch.”

But he noted that significant time had lapsed since the last test flight.

Nasa is aiming in 2027 to test an in-orbit rendezvous between its spacecraft and at least one lunar lander, which both SpaceX and rival Blue Origin – the Jeff Bezos-owned firm – are racing to develop.

That Artemis phase is meant as a step towards carrying out a crewed lunar landing before the end of 2028, and before the end of Donald Trump’s presidency.

But for Swope, “there is a long way to go and many more test flights before Starship is ready for the next Artemis mission.”

Ahead of Friday’s test, Nasa Administrator Jared Isaacman appeared during the pre-launch SpaceX program and said: “We’re looking forward to seeing this fly, because hopefully at some point in the not-too-distant future we’re going to join up in Earth orbit.”

Following the test, Isaacman posted praise on X, congratulating SpaceX on “a hell of a V3 Starship launch.”

“One step closer to the Moon...one step closer to Mars,” the Nasa official said. — AFP

KUALA LUMPUR, May 23 — Every year, aspiring student coders from all over the world join the Apple Swift Student Challenge. 

From the best entries, a selected few get not just a certificate and cool prizes but an invitation, travel and board provided, to Apple Park for the Worldwide Developers Conference.

These 50 winners are dubbed Distinguished Winners whose entries have been considered “truly exceptional” by Apple and this year, Malaysian Jasmmender Kaur is among them.

Ipoh-born Jasmmender, 22, is currently pursuing her bachelor’s in computer science at Taylor’s University Malaysia, specialising in data science.

She said her interest in technology started at an early age: “My journey with technology actually started at home at the age of 10 when a curiosity about how things worked led me to my first computer.”

After picking up the basics of computing at school, she soon started using HTML and CSS to code her own websites as well as built mini-games in Scratch.

By the time she was 15, she had also created a tuition centre management system and with Python, created a simple quiz game.

As to how she became acquainted with Apple’s developer ecosystem, she said, “I actually taught myself Swift independently through YouTube tutorials and Apple’s official developer tutorials, documentation and workshops.”

Apps are in themselves solutions, solutions to specific problems and what Jasmmender saw as a challenge was to help people understand what’s actually going on behind AI.

Her app, Unveil, is meant to educate users about how AI works, saying that the name came from what it does, “pulls back the curtain on the inner workings and the technical backgrounds of AI, but in a very friendly and approachable manner.”

With AI already being used in a myriad of ways, at scale, Jasmmender said, “But most people have no understanding of how or why these decisions are being made.”

Jasmmender said that existing information is not tailored as much for the everyday person, either being too formal or targeted at those already in tech. 

“I felt like in Malaysia, AI adoption is growing, but openness without proper understanding is a very dangerous gap,” she said.

She first listed down what she knew about AI, then noted things that she felt would be important for people to understand and know about.

With that base, she then used AI tools like Claude to create a structure and figure out how she could create something that could be translated easily for everyone.

She added, “But I also got advice from my friends as well about things that they wish they knew before they got into AI.”

Jasmmender considered her win as invaluable in many ways.

“It not only gave me the experience and the lessons, but it opened doors to connect with incredible developers and professionals globally, especially through the network from the Apple community,” she said.

It also allowed her a platform to advocate for AI literacy on a much larger scale, something that she is deeply passionate about.

“Everyone deserves to understand the system that’s shaping their world,” she said.

Apple’s Worldwide Developers Conference will take place in Apple Park at Cupertino, California from June 8 to 12.

Caption: Jasmmender Kaur learned to code from an early age and is now in her final year studying computer science. — Picture courtesy of Apple

MONTREAL, May 22 — World Cup fans are wielding artificial intelligence to mass-produce viral songs supporting their teams ahead of next month’s tournament.

As the fan-made football anthems are raking in millions of plays across YouTube, TikTok and Instagram, experts say that the viral tunes raise questions about song ownership, artist compensation and the valuation of human creativity.

But many users do not appear to mind, with some even showing a preference for the AI-generated songs over an official anthem football’s world governing body Fifa commissioned from musicians Jelly Roll and Carin Leon.

A highly-anticipated World Cup track from Shakira was also released last week, but the fad of AI fan songs was still drumming up excitement on social media for the tournament taking place in cities across the United States, Canada and Mexico in June and July.

The trend appears to have started with a song dedicated to the French team, Imbattables, released in February by artist Crystalo, who is listed on Spotify as France’s “premier AI musical creator.” 

The song begins with a call-and-response listing the names of Kylian Mbappe and other star French national players.

A Brazilian anthem followed with a similar name-chanting format and a trending phonk melody that producer Guilherme Maia, who goes by the artist name M4IA, said he created by layering together different elements he had put together with the help of AI.

Tracks for top sides Portugal, Argentina and Germany, as well as many others, soon sprang up across platforms, garnering more praise from fans.

But while the Brazilian version closely resembled the French prototype, the later songs copied Maia’s format exactly. 

Each recycled the phonk beat and listed players’ names before calling for respect for the squad’s “king” — a feature reserved for the likes of Cristiano Ronaldo in the Portugal tune or Lionel Messi in Argentina’s version.

“What I see happening now is more about people following a trend or trying to recreate a feeling,” Maia told AFP, saying that artistic emulation has always existed in music.

While he was enthusiastic about the possibilities AI opened up for production, he acknowledged that the technology raises new questions about authorship and copyright.

“In music, there are clear rules. You can’t just copy someone else’s work or use samples without permission, even if AI is involved.”

Lack of credit 

Maia stressed that he built the track on his own and used AI as an assistant when creating certain elements, rather than asking a music generation tool like Suno to create a song with one prompt.

But, Jason Palamara, an assistant professor of music technology at Indiana University, said that with the way the models exist, there is a lack of clarity over how artists are credited if their copyrighted work is used to train them.

“It had to come from somewhere,” he said.

The inconsistencies that can appear in the AI-generated images can also pop up in music created with the technology.

For example, a fan-made World Cup song for Portugal was sung with a Brazilian accent, while a Colombian version read James Rodriguez’s first name with an English rather than Spanish pronunciation.

Music created with AI can also lack complexity, Palamara said.

“It’s one compact product, rather than a product where there’s multiple tracks that have gone into it, where it has more texture.”

Still, Morgan Hayduk, co-CEO of music rights software company Beatdapp, said that listeners enjoying the World Cup fan songs may not be seeking artistic complexity.

“There seems to be a cohort of people who actually don’t care,” Hayduk observed. “They like the music, and they like the back story that it came from a large language model and not a songwriter or a group.”

He said that despite concerns over how the industry will adapt to AI, quick-fix songs that can be chanted by fans or featured in advertisements are a clear use-case for AI-generated music in its current stage.

“Knowing what goes into a generative output, like a World Cup fan song, is the thorny Rubicon that the music industry has to cross now.” — AFP 

NEW YORK, May 22 — Spotify is teaming up with Universal Music Group to let users create AI-powered remixes and covers of songs by artists on the label — for an extra fee on top of a standard subscription.

The new feature, announced Thursday, will only apply to artists who have given their consent, and both the original performer and songwriter will receive a share of any revenue generated.

“For the first time, fans will be able to legally create covers and remixes from participating artists and songwriters’ catalogs, with both the original artist and the songwriter sharing in the value created,” said Charlie Hellman, Spotify’s head of music, speaking at the company’s investor day.

Until now, Spotify had banned AI-generated music derived from a specific artist’s work without their express authorisation, though it does allow AI music to be uploaded more broadly, including content associated with artists whose image and identity were generated using artificial intelligence.

The deal puts Spotify in direct competition with Suno and Udio, the two dominant AI music apps on the market, both of which have been posting strong growth.

After initially allowing users to create AI music without agreements in place with artists or labels, both startups have shifted course in recent months.

Udio has signed deals with Universal Music Group and Warner Music Group, which also reached an agreement with Suno.

UMG and Sony are nonetheless still in litigation with Suno, with proceedings ongoing before a US federal court in Massachusetts.

Hellman said the feature would give artists and songwriters “a brand new source of income on top of what they already earn on Spotify.”

Universal Music Group CEO Lucian Grainge called the initiative “firmly artist-centric, rooted in responsible AI,” saying it would “drive growth for the entire ecosystem.”

Also announced at the investor day, Spotify said it would give paying subscribers early access to concert tickets for their favorite artists before they go on general sale.

The new service called “Reserved” is launching in the United States this summer before expanding to other markets.

Subscribers will be selected based on listening data — including how often they stream a given artist, the range of tracks they play from that artist’s catalog and whether they have saved songs to their library.

They will have around 24 hours to purchase up to two tickets through a partner ticketing platform, the company said.

Spotify said the initiative would help route tickets to genuine fans rather than scalper bots, which have long frustrated both concertgoers and artists. — AFP 

NEW YORK, May 22 — SpaceX’s blockbuster IPO filing included some out of this world details, including a provision that founder Elon Musk’s massive bonus only kicks in if one million humans settle on Mars.

The bonus structure, laid out in SpaceX’s prospectus filed with US regulators Wednesday, reads less like a compensation agreement and more like a science fiction plot.

Musk’s bonus is contingent on SpaceX’s stock market value hitting targets ranging from US$400 billion (RM1.58 trillion) to US$6 trillion — along with the company moving a million people to a planet 140 million miles (225 million kilometers) away.

Musk describes that ambition as essential to the long-term survival of the human race, though most experts see it as being at least decades away.

Still, Musk will do just fine if the IPO goes ahead in the coming weeks as planned.

At the company’s reported target valuation of US$1.75 trillion, Musk’s existing stake would be worth an estimated US$735 billion — before a single person sets foot on the Red Planet.

A second, smaller bonus ties an additional 60 million shares to a different moonshot: building data centres in orbit capable of delivering 100 terawatts of computing power per year — a figure that dwarfs anything that exists on Earth today.

SpaceX filed for its long-awaited IPO Wednesday, targeting a listing on the Nasdaq stock exchange under the ticker “SPCX” in what could be the largest public offering in Wall Street history.

The company’s Starship rocket — whose latest iteration could launch Thursday — is explicitly designed with Mars colonisation in mind. — AFP 

SAN FRANCISCO, May 21 — Airbnb, facing tighter local regulations on short-term home rentals, announced Wednesday it is adding boutique hotels, car rentals and grocery delivery to its app in a bid to transform itself into a one-stop travel shop.

The rollout marks the latest step in Airbnb’s push to capture more of the travel spending that currently flows to competitors like Booking.com and Expedia.

This evolution — 18 years after the company’s scrappy beginnings in San Francisco — is one of Airbnb’s answers to increasingly tough restrictions on short-term rentals in key markets.

In December, Spain hit the company with a €65 million (RM299.3 million) fine over more than 65,000 non-compliant listings, and Barcelona decided not to renew thousands of rental licenses when they expire in 2028.

New York has banned nearly all short-term private rentals since 2023, and Paris stepped up its crackdown on illegal listings in 2026.

“Travel shouldn’t just be convenient. It should be meaningful,” CEO Brian Chesky said in a statement. “The best trips help you explore, learn, and come home a little different than when you left.”

The updated app adds grocery delivery through Instacart in more than 25 US cities, as well as airport and train station transfers and luggage storage services in more than 160 cities worldwide.

The platform will also offer car rentals, though the company has not yet named its partners.

Chesky outlined the new features at a presentation at the company headquarters in San Francisco.

“We know that a home is not right for every type of trip, like if you’re booking last minute, you’re staying for just one night, you’re on a quick business trip, or when you’re searching for an Airbnb, like in New York, and none are available....Boo,” he said, joking about the restrictions in the Big Apple.

“In this case, the best option might just be...hotel! That’s right, I said it,” Chesky added, as Airbnb employees clapped.

Chesky argued that boutique and independent hotels account for 60 per cent of hotel rooms around the world and his company wants to team up with them.

“They want an ally, they want a platform designed just for them, so they can compete with the chains,” Chesky said.

The app is also getting a range of artificial intelligence features, including a virtual support assistant available in 11 languages.

Airbnb posted revenue of US$2.68 billion in the first quarter of 2026, up 18 per cent from a year earlier. — AFP 

LONDON, May 21 — TikTok and Alphabet’s YouTube have failed to set out meaningful steps to protect British children from harmful online content, media regulator Ofcom said on Thursday, citing data showing widespread exposure on their platforms.

Ofcom said neither company had made significant new commitments to make recommendation feeds safer, despite evidence that these feeds are the main route through which children encounter harm.

Governments are stepping up efforts to improve child safety online, with Prime Minister Keir Starmer urging social media firms to take greater responsibility.

Britain is consulting on tighter curbs, including a possible ban on under-16s using social media, modelled on Australia’s landmark move, to tackle what it calls addictive design features.

Systems still fall short, Ofcom says

Ofcom research found 73 per cent of 11- to 17-year-olds were exposed to harmful content over four weeks, mainly via personalised feeds. TikTok was cited most often, followed by YouTube, Meta’s Instagram and Snap’s Snapchat.

Ofcom said TikTok and YouTube maintained their existing systems were sufficient, but the regulator said its evidence suggested their feeds “are still not safe enough”.

A YouTube spokesperson said: “YouTube provides industry-leading, age-appropriate, high quality experiences for young viewers, working with child safety experts to deliver protections that support millions of families across the UK.

“We welcome today’s news that others across the industry will soon adopt similar features.”

A TikTok spokesperson said it was “very disappointing that Ofcom has failed to acknowledge both our longstanding and newer safety features.”

“We will continue to make ongoing investments in safety measures for our users,” the person added.

Snap, Meta and Roblox have all agreed to introduce stronger protections against online grooming following Ofcom’s demands last month.

Under those commitments, Snap will block adult strangers from contacting children by default and expand age checks in Britain. Meta plans new controls on teen accounts and AI tools to detect suspicious conversations. Roblox will allow parents to disable direct messaging for under-16s.

Government urged to strengthen law

Nearly a year after new child safety duties under the Online Safety Act took effect, Ofcom said there had been little overall improvement in children’s exposure to harmful content.

The regulator said major platforms dominate children’s online activity, with YouTube used by 67 per cent of children and TikTok by 60 per cent, and 95 per cent using at least one social media or video-sharing service.

It also flagged weak enforcement of minimum age rules, noting 84 per cent of children aged eight to 12 use services requiring users to be at least 13.

Ofcom said current legislation does not clearly require companies to keep underage users off their platforms and urged the government to strengthen the law.

Separately, Elon Musk’s X has agreed to step up enforcement against illegal hate speech and terrorist content, including reviewing such material within 24 hours on average and sharing quarterly data with the regulator.

Ofcom is still examining X’s systems and its Grok chatbot after a Reuters report earlier this year found the tool could generate sexualised images in many cases despite user warnings. — Reuters 

WASHINGTON, May 21 — Speakers promoting AI are getting booed at universities, voters are rebelling against data centres, and even AI-friendly Trump administration officials are starting to retreat as an artificial intelligence backlash gathers pace across the United States.

The rapid spread of the emerging technology is seeing early enthusiasm give way to concerns about unemployment, rising costs, misinformation and security.

“People are thinking about what their future is going to look like. That existential fear is a very animating anxiety,” said Christabel Randolph, acting executive director at the Centre for AI and Digital Policy, a Washington-based think tank.

Former Google CEO Eric Schmidt got a taste of that feeling Friday when he was delivering a graduation speech at the University of Arizona.

Wearing a black academic gown and a tassel-topped cap, Schmidt urged students not to fear the AI-fuelled technological transformation that he said will “touch every profession, every classroom, every hospital, every laboratory, every person and every relationship you have.”

Instead of clapping, his speech prompted loud boos.

Accept change ‘or pay’ 

According to opinion polling cited by the Semafor news outlet, 70 per cent of Americans think AI is moving too fast, over 50 per cent have negative views of it, and just 18 per cent of young people feel hopeful about it.

With the US economy battered by stubborn inflation and the tech industry seeing AI-fuelled layoffs, young Americans fear their costly university degrees, many paid for with large student loans, will be rendered useless by AI, leaving them without jobs and pay.

When Scott Borchetta, CEO of Big Machine Records, tried to tell Middle Tennessee State University graduates to embrace inevitable change, he too got a hostile reception.

“You can hear me now or you can pay me later,” he quipped. “Do something about it, it’s a tool, make it work for you.”

Booing followed.

‘Really, really angry’ 

AI expansion is driving a massive build-out of data centres -- and that infrastructure is now becoming a political flashpoint.

Data centres consume large amounts of electricity and can raise utility costs, which has seen local officials supporting AI projects suffering losses at the ballot box in recent months. Some of the discontent has spilled into violence.

Last month, a young man threw a Molotov cocktail at the California home of OpenAI CEO Sam Altman. In a separate incident a few days earlier, a city council member in the state of Indiana had his door struck by gunfire after he expressed support for a data centre construction project.

A note the attackers left under his doormat read “No Data Centres.”

Randolph, the AI expert, cited a May Gallup poll showing that AI data centres are even less popular than nuclear power plants, with 71 per cent of Americans opposing local AI data centres compared with 53 per cent opposed to nearby nuclear plants.

“Americans are really, really angry and upset about AI data centres because of the noise, the pollution, the impact on their electricity bills, on water supplies,” she told AFP, adding that AI expansion will be a key issue in the November midterms and possibly in the 2028 presidential vote.

“It’s becoming a very relevant political issue,” she added.

‘Have to be careful’ 

The Trump administration itself appears to be changing its stance.

Since returning to the White House in 2025, President Donald Trump has positioned himself as an advocate for rapid AI development, rolling back Biden-era safety requirements and dismissing regulation as a constraint on US competitiveness with China.

But in recent months, the administration announced that it wants to vet AI models before they are released, urged Congress to adopt nationwide regulations on AI and discussed AI guardrails with China.

Asked about the risks of AI on Fox News’ Mornings with Maria programme last month, Trump answered:

“There are a lot of good things, but we have to be careful with it.” — AFP

MOUNTAIN VIEW, May 20 — Google on Tuesday showed off new plans to turn its famous search bar into an AI assistant that can book restaurants, track news and contact businesses – just by asking a question.

After three years of struggling to keep up with ChatGPT, Google is racing to roll out artificial intelligence tools that build on its grip over online search.

The company’s Gemini AI app now has 900 million monthly users, twice as many as last year. Its AI-powered search feature, AI Mode, is also taking off, with a claimed one billion monthly users worldwide.

On Tuesday, at Google’s annual developer conference near its California headquarters in Mountain View, CEO Sundar Pichai unveiled the next step: Gemini Spark, a personal AI agent available starting next week for top-tier subscribers in the United States.

Google’s search engine will also get a new upgrade for US users this summer: always-on AI agents that can alert you to news, book activities, and manage shopping lists.

The changes to Google search, which the company said were its biggest in 25 years, will also see a widened search box to make room for more complicated queries people use for chatbots.

“I love how search has become less about individual queries and feels more like an ongoing conversation, giving users deeper insights and connecting you with the vastness of the web,” Pichai told journalists.

Many of the features ride a wave of “agentic” AI that has gripped Silicon Valley since Austrian developer Peter Steinberger in late 2025 launched OpenClaw – a platform that lets AI book flights, manage emails and build apps from chat prompts.

OpenAI hired OpenClaw’s creator earlier this year and the tech giants are now racing to bring agentic features to mainstream users, despite security concerns and the soaring computing costs that come with them.

To stay ahead of rivals Anthropic and OpenAI, Google on Tuesday also launched the latest version of its AI model, Gemini 3.5 Flash.

Google says it runs four times faster than top competing models – including Anthropic’s Claude Opus and OpenAI’s ChatGPT 5.5 – while performing at a similar level.

The model is now the default across the Gemini app, AI Mode search and other Google services. A more powerful version, Gemini 3.5 Pro, is expected next month.

Google also announced it was teaming up with OpenAI on one front: to help stop the spread of fake or manipulated content, the ChatGPT maker has adopted SynthID, Google’s tool for invisibly watermarking AI-generated images.

End of clicks?

Google’s growing AI features could spell trouble for news websites and online publishers.

By keeping users inside its own apps and tools, Google makes it less likely that people will click through to outside websites – cutting into their traffic and ad revenue.

Google searches already end 58 per cent of the time without users clicking on any website, according to a lawsuit filed against the company by Penske Media, which publishes the Hollywood Reporter and Rolling Stone.

In Europe, a major publishers’ group has complained to the European Commission, saying Google uses news content to fuel its AI summaries without paying for it.

France is the only major European country where AI Mode is still unavailable, and remains at the centre of a bitter fight between Google and French publishers.

However, Google’s legal troubles are not limited to Europe.

A US court found it guilty of illegally monopolising online search in 2024, and the company could still be forced to break up parts of its business.

The Justice Department in February appealed a ruling that had stopped short of making Google sell its Chrome browser.

A hearing is not expected until the end of the year at the earliest, or possibly 2027. — AFP

LOS ANGELES, May 16 — Alphabet’s YouTube and Snap have reached settlements in the first case set for trial in litigation seeking to force social media platforms to cover the costs school districts incur to combat a youth mental health crisis they say the companies fuelled.

The settlements were detailed in court filings yesterday in federal court in Oakland, California, and resolve claims by a Kentucky school district that is still due to take Facebook and Instagram parent Meta Platforms and TikTok to trial on June 15.

Terms of the settlements with Breathitt County School District in rural Eastern Kentucky were not disclosed.

“This matter has been amicably resolved and our focus remains on building age-appropriate products and parental controls that deliver on that promise,” a YouTube spokesman said in a statement.

Snap, the parent company of Snapchat, did not respond to a request for comment.

More than 3,300 lawsuits involving addiction claims are pending in California state court against the social media companies. Another 2,400 cases brought by individuals, municipalities, states and school districts have been centralized in California federal court.

In a landmark trial, a Los Angeles jury on March 25 found Meta and Alphabet’s Google negligent for designing social media platforms that are harmful to young people. It awarded a combined US$6 million (RM24 billion) to a 20-year-old woman who said she became addicted to social media as a child.

The companies have denied the allegations and say they take extensive steps to keep teens and young users safe on their platforms.

Breathitt is one of more than a thousand school districts suing the social media companies over claims they caused a mental health crisis among students and then saddled schools with the fallout.

The school district has been seeking over US$60 million to cover the costs of counteracting social media’s impact on students’ mental health and to fund a 15-year mental health program to abate the problem.

It also seeks a court order requiring the companies to modify their platforms to reduce addictive features.

Its case is a bellwether, or test case, for over a thousand similar school districts’ lawsuits.

Judges and attorneys often use bellwether verdicts to assess the potential value of remaining claims and guide settlement talks. Typically, several bellwether cases are tried before reaching a broader resolution. — Reuters

KUALA LUMPUR, May 14 — Meta has announced a new feature called “Incognito Chat with Meta AI”, which it claims enables completely private conversations with AI on WhatsApp and the Meta AI app.

According to Meta, the new feature is built on its “Private Processing” technology and is designed to ensure that conversations remain inaccessible to anyone else, including Meta itself.

The company said current “incognito-style” AI modes offered by other apps may still allow providers to see incoming questions and outgoing responses.

In contrast, Meta claims its Incognito Chat is designed so that no one can read your conversation, not even Meta themselves.

Meta added that when users start an Incognito Chat with Meta AI, the conversation becomes a private temporary session that only the user can access. Messages are processed in what Meta describes as a secure environment that even the company cannot access.

The company plans to expand Private Processing to additional AI features in the coming months. — SoyaCincau pic

By default, conversations are not saved and messages disappear automatically after the session.

Meta said the feature is aimed at users who may want to ask sensitive questions or share private information involving finances, personal matters, health, or work-related data with AI tools.

The company plans to expand Private Processing to additional AI features in the coming months.

This includes a new “Side Chat” feature for WhatsApp, which would allow Meta AI to privately assist users based on the context of an ongoing chat without interrupting the main conversation.

Meta said Incognito Chat with Meta AI will roll out gradually on WhatsApp and the Meta AI app over the coming months.

You can learn the technical details about Private Processing from Meta’s technical whitepaper published via its AI research portal. — SoyaCincau 

KUALA LUMPUR, May 14 — Acer Malaysia has launched a fresh wave of gaming hardware, unveiling new laptops, desktops, monitors and portable displays designed to cater to gamers at every level. 

The rollout is accompanied by the ongoing “Predator Kasi Onz Roadshow” at Sunway Pyramid, where customers can score instant discounts, eWallet credits and free gifts.

Announced are the Predator Helios Neo 16 gaming laptop, Predator Orion 5000 desktop, Predator XB273U X2 monitor and the Nitro PG241Y P1 portable display — offering gamers more flexibility across mobile, stationary and multi-screen setups.

Leading the pack is the Predator Helios Neo 16, a mid-tier gaming laptop aimed at socially connected gamers who want solid performance without breaking the bank. 

Available in multiple configurations, it packs up to an Intel Core i7 processor and NVIDIA GeForce RTX 5070 laptop GPU, paired with a 5th Gen AeroBlade 3D fan to keep thermals in check during marathon sessions.

The 16-inch 16:10 IPS display reaches up to WQXGA+ resolution, with a 180Hz refresh rate, 3ms response time and 100 per cent sRGB colour coverage. 

Connectivity is handled by Intel Killer DoubleShot Pro (Wi-Fi 6E and 2.5G Ethernet), while DTS:X Ultra Audio and Acer PurifiedVoice with AI noise reduction aim to keep comms crisp.

Prices start from RM7,299 for RTX 5050 variants, topping out at RM9,999 for the RTX 5070 model.

For those who prefer a stationary command centre, the Predator Orion 5000 arrives in a 45L chassis made with 65 per cent PCR plastic, featuring an EMI‑compliant tempered glass side panel and customisable ARGB lighting. 

Under the hood, an Intel Core Ultra 7 Processor 265F and NVIDIA GeForce RTX 5070 graphics are kept cool by the Predator CycloneX 360 cooling system.

It comes with 32GB DDR5 6000 XMP MHz RAM (upgradable to 128GB) and 1TB SSD storage, plus Wi-Fi 7 support. The desktop is priced at RM14,999.

Acer also unveiled the Predator XB273U X2, a 27-inch QHD gaming monitor with an overclocked 210Hz refresh rate and 0.5ms response time, aimed at reducing motion blur in fast-paced titles. 

It retails for RM799 and includes a three-year pick‑up and return warranty.

Meanwhile, the Nitro PG241Y P1 is touted as Malaysia’s first 23.8‑inch portable gaming monitor, featuring a 144Hz IPS panel that connects via a single USB‑C cable. 

A built‑in kickstand allows for easy placement on desks, mounting or even business meeting screens. It is priced at RM649.

All products are available immediately at the Acer eStore, official Shopee and Lazada stores, and authorised resellers nationwide. 

However, the biggest savings are at the Predator Kasi Onz Roadshow at Sunway Pyramid Blue Atrium, running from May 14 to 17, 2026.

For more details, visit Predator Malaysia’s Facebook page (@PredatorGamingMy) or call Acer’s Product Infoline at 1800-88-1288 (9am–6pm, weekdays).

New York, May 12 — Before he opened fire on the Florida State University campus last year, killing two people and wounding six others, Phoenix Ikner had a conversation.

Not with a friend, a parent or anyone who might have talked him out of it — but with an AI chatbot.

According to evidence gathered by Florida’s attorney general, the student had asked ChatGPT which weapon and ammunition would be best suited for his attack, and when and where he could inflict the most casualties.

The chatbot, investigators say, answered his questions.

Now Attorney General James Uthmeier wants to know whether that makes OpenAI a criminal.

“If the thing on the other side of the screen was a person, we would charge it with homicide,” he said, announcing a criminal investigation into ChatGPT maker OpenAI and leaving open the possibility of charges against the company or its employees.

The case surrounding the April, 2025 shooting has thrust a provocative question into the legal spotlight: Can the creators of an artificial intelligence be held criminally liable for the role their AI played in a crime — or even a suicide?

Legal experts say it’s a realistic, if deeply complicated, proposition.

Criminal product?

Criminal prosecutions of corporations are possible under US law, though they remain relatively uncommon.

Late last month, Purdue Pharma was hit with more than US$5 billion in criminal fines and penalties for its role in fuelling the opioid crisis.

Volkswagen was previously found guilty in the emissions cheating scandal, Pfizer over its promotion of the anti-inflammatory drug Bextra and Exxon for the Exxon Valdez oil spill in Alaska.

But those cases all involved human decisions — executives, salespeople or engineers who made choices and cut corners.

The Ikner case is different, and that difference is precisely what makes it so legally treacherous.

“Ultimately, it was a product that encouraged this crime, that did the act of the crime,” said Matthew Tokson, a law professor at the University of Utah. “That’s what makes this case so unique and so tricky.”

Legal experts consulted by AFP say the two most plausible charges would be negligence or recklessness — the latter involving a deliberate choice to ignore known risks or safety obligations.

Such charges are often treated as misdemeanours rather than felonies, meaning lighter sentences if convicted.

The bar, however, is high.

“Because this is such a frontier issue, a more compelling, more clear-cut case would probably involve internal documents recognising these risks and maybe not taking them seriously enough,” Tokson said.

“In theory, you could get liability without it,” he said. “But in practice, I think that’d be difficult.”

In criminal law, “the burden of proof is higher,” noted Brandon Garrett, a law professor at Duke University — with prosecutors required to establish guilt beyond a reasonable doubt.

OpenAI, for its part, insists ChatGPT bears no responsibility for the attack.

“We work continuously to strengthen our safeguards to detect harmful intent, limit misuse, and respond appropriately when safety risks arise,” the company said.

Civil or criminal?

For those seeking accountability, a civil lawsuit may offer a more viable path.

Such an approach might push companies to design their products more carefully — or at least force them to reckon with the human cost of getting it wrong, said Tokson.

Several civil cases have already been filed against AI platforms in the US — many involving suicides — though none has yet resulted in a judgment against the companies.

In December, the family of Suzanne Adams sued OpenAI in California court, alleging that ChatGPT contributed to the murder of the Connecticut retiree by her own son.

Newer versions of ChatGPT have introduced additional safeguards, acknowledged Matthew Bergman, founding attorney of the Social Media Victims Law Center.

“I’m not saying that they are adequate guardrails, but there are more guardrails in effect,” he said.

A criminal conviction, even with a modest sentence, could still inflict serious damage, including a “big reputational impact,” Tokson said.

But for Garrett, prosecutions — however dramatic — are no replacement for the regulatory frameworks that Congress and the Trump administration have so far failed to put in place.

That, he said, would be “a much more sensible system.” — AFP

TOKYO, May 8 — Japanese gaming giant Nintendo said Friday it will hike the price of its Switch 2 gaming console as memory chip costs soar, warning that net profit would fall 27 per cent this year.

Sony — whose PlayStation5 has already risen in price — was more upbeat, projecting a 13-per cent rise in income but still with falling sales of its ageing console.

Nintendo said the Switch 2 price in Japan will rise 20 per cent from May 25, and from September 1 by 11 per cent in the United States to US$499.99 (RM1,960) and in Europe by six per cent to €499.99.

For the year to next March, Nintendo expects net profit to drop 27 per cent to ¥310 billion on sales of ¥2.05 trillion, marking a fall of 11.4 per cent.

It also forecast ¥370 billion yen in operating profit, considerably below the average analyst estimate of ¥480 billion, according to Bloomberg News.

Net profit surged 52 per cent to ¥424 billion last year on annual sales of ¥2.31 trillion, nearly doubling from the previous year, Nintendo said in a statement.

“Nintendo Switch 2 got off to a good start following its launch in June and global sales continued to grow after that,” the company said.

It sold 19.86 million units of the new console by March, thanks to games like Pokemon Pokopia, Mario Kart World and Donkey Kong Bananza.

Memory chips 

Price rises of memory chips fuelled by the artificial intelligence boom have hit makers of games consoles, smartphones and other devices, while disruptions linked to the Iran war have exacerbated supply problems.

Sony said Friday that it sold 16 million PlayStation5 units in the past fiscal year, down from 18.5 million in the previous 12 months.

With 92 million PlayStation2 units sold since its launch in 2020, analysts said the firm was well placed to benefit from the release of smash hit Grand Theft Auto VI, due in November.

“If there is a game that can sell PlayStations by the millions, it is this one,” Gaming industry consultant Serkan Toto told AFP.

For the year to March 2027, the game division is expected to enjoy higher profits despite falling sales, Sony said.

“Sony’s more mature PS5 console cycle leaves it better placed to weather higher memory costs,” said Amir Anvarzadeh, strategist at Asymmetric Advisors.

“Having already moved past the heavy hardware penetration costs typical of earlier years, Sony’s bottom line stands to benefit significantly from the high-margin software sales and ecosystem engagement this launch should trigger,” Anvarzadeh said.

Nintendo though is in a more difficult position, Toto said, as Switch 2 customers are “especially price sensitive”.

“The first year game lineup for Switch 2 is much weaker than for its predecessor,” he said.

“But now it’s time for them to really step on the gas on the software side.” — AFP 

KUALA LUMPUR, May 8 — Once upon a time (or when the pandemic happened) I briefly considered podcasting.

Then I realised I don’t really like podcasters.

I tell you this story to explain why I have more headphones than the average person.

There’s a set of planar magnetic headphones (to hear more detail in my music), a neutral-pair of headphones (for mixing audio), a wireless gaming headset because my other headsets were too awkward to use with my PS5, the previous AirPods Max (a review unit) as well as wired IEMs and of course AirPods. 

For a lot of people, long ago, whatever earphones you got with your phone was enough and if they had a mic, that was a bonus.

With the popularity of streaming services and the rise of Chi-fi (Chinese audio tech that offers great sound cheaper than ever before) a lot more people care about how their audio sounds but the question must be asked, who is the AirPods Max 2 for?

Same old, same old

I usually take photos of review items before I take them for a test spin, but I decided to just go around with the new AirPods Max and then take photos.

Reader, I was an idiot.

Sturdy build and design does make up for drawbacks some other headphones have. — Picture by Erna Mahyuni

There were ugly stains on the AirPods Max headband as well as the earcups thanks to that annoying thing called sweat.

That’s what I don’t get about how people were wearing these everywhere, in public. I tried walking with them to the train station, around in malls, visiting cafés and while they weren’t uncomfortable to wear, sweat did get on them. 

Admittedly, the staining wasn’t something a little time with a wet wipe couldn’t fix. 

While some audiophiles hated the mesh on the headband and cups, and would have preferred the leather or faux leather of more traditional headphones, at least they were easy to clean. 

I can snap off the earcups (Apple calls them Ear Cushions), wipe them off, air dry them and I don’t have to worry about the peeling I’m already seeing on my other headphones.

Apple also sells replacement AirPods Max Ear Cushions though there are OEM ones being sold for cheaper. 

They’re still somewhat heavy which makes them less ideal for going around in, especially in an age where retro ultra-light (and ultra-flimsy) headphones are popular again.

Design-wise there are trade-offs and some people might find these more comfortable than others, depending on head and ear size.

If you’re lefthanded all the controls being on the right side might take some getting used to. — Picture by Erna Mahyuni

Sound: A lot of fun, especially for modern music

If you listen to music mostly on Apple Music then the AirPods Max 2 is hard to beat.

Should you listen to a lot of vocal-heavy music, you might enjoy how these cans put the vocals rather forward, something I noticed (and liked) when listening to Zara Larsson’s Blue Moon remix with Kehlani and that classic neo-soul number Tell Me from Groove Theory.

Apple’s decided to tone it down a little with the bass, which I consider a good thing since these aren’t Beats headphones.

Instead you get a lot more openness, better detail and an expansiveness you can appreciate with songs like Zhou Shen’s 大鱼  that puts Zhou’s airy countertenor at the centre but you can still feel the deep rumble of percussion and the swirling, flowy string arrangement.

Put on Luther Vandross’ Superstar/Until You Come Back to Me and you can hear in crisp detail, the opening strings and flute, then the steady hum of the keys and that slight rasp, the lush, warmth of Vandross’ honeyed tone.

If there’s one song I feel the AirPods Max 2 was made for, it’s this one.

What else are you getting?

Apple has tweaked the Bluetooth wireless experience by making it super-easy to switch between full ANC, Transparency mode, Adaptive (the Goldilocks option between ANC and Transparency mode), all with the controls on the right Ear Cushion.

The AirPods Max 2 are good for easy pairing (most of the time) to Apple devices but are unlikely to take over other niche headphone uses. — Picture by Erna Mahyuni

If you’re left-handed that’s perhaps one mild annoyance but not a big one.

You also can use the Digital Crown to take photos, that I guess makes selfies easier (especially if you want to show off your AirPods Max 2).

Performance-wise, the new headphones has an updated processor in the H2, the new model also supports extra features such as Live Translation, improved Spatial Audio, Voice Isolation and nine microphones, eight for ANC and three for detecting voices (two of the three are shared with ANC and the third is an extra one). 

You’re still getting the weird AirPods Max case that isn’t particularly protective but helps tell the AirPods Max 2 they’re supposed to be in Auto-sleep mode because Apple still won’t give us a power button for it.

There is no weather-resistance or water-resistance here so do not wear them to the gym or if there’s any chance for rain; for exercise or walks in the park, you’re better off with AirPods. 

Take them home?

If you plan to use them with an Android phone, best not.

The AirPods Max 2 were optimised for use with Apple devices, whether iPhones, Macs or the Apple TV. 

If you have the previous model, upgrade if you want the Type-C port as well as the improved Adaptive Audio experience but if you mostly use it at home and the sound still makes you happy, you can probably hold onto it till it dies. 

The AirPods Max 2 are already available at official Apple retailers and online for RM2,299.

SAN FRANCISCO, May 8 — Generative artificial intelligence is being used by 17.8 per cent of the world’s working-age population, but the gap between wealthy and developing nations continues to widen, according to a report published Tuesday by Microsoft.

In the first quarter of 2026, 27.5 per cent of people aged 15-64 in developed countries used a generative AI tool, compared with 15.4 per cent in the developing world — a gap that widened by 1.5 percentage points from the second half of 2025, according to the report’s estimates.

The divide stems from significant inequality in access to internet connectivity, basic digital skills and electricity, according to the Microsoft AI Economy Institute.

AI model performance — historically stronger in English as most of the major AI companies are based in the US — is also slowing the spread of such tools in non-English-speaking countries.

But progress in processing non-European languages is fuelling a catch-up in adoption in some countries, particularly in Asia, the US tech giant noted.

The United Arab Emirates tops the ranking of AI usage at 70.1 per cent, followed by Singapore, Norway, Ireland and France.

The estimates were based primarily on measurements from computers running Windows and Microsoft products such as Bing and Copilot.

They only partially captured usage on Apple devices, and consolidated data was lacking for Russia, Iran and China.

The United States — home to dominant large AI models like ChatGPT, Claude and Gemini — ranked only 21st, at 31.3 per cent.

AI usage in China — the world’s second-largest economy which is jostling with the US for an edge in the AI race — was 16.4 per cent, the report said.

Pushing back against fears of job losses driven by automation, Microsoft argued in the report that AI coding tools “could increase demand for developer jobs.”

The company cautioned, however, that “it is still too early to know the full impact” of AI on the labor market.

For the first time in its history, the company itself offered voluntary departures to nearly 9,000 of its US-based employees in April.

According to Layoffs.fyi, a private aggregator, nearly 99,000 people have been laid off in the tech sector since January 1, primarily in the United States. — AFP 

KUALA LUMPUR, May 3 — TNG Digital has announced that it will discontinue Quick Payment and Auto Reload features for debit cards on the TNG eWallet starting 13 May 2026.

The change affects users who rely on debit cards for automatic top-ups and seamless payments. TNG eWallet users who currently use credit card for Quick Payment and Auto Reload are not affected.

According to the official statement, the feature had been placed under maintenance as part of an ongoing review of transaction performance and reliability. Following the review, TNG Digital decided to permanently discontinue support for debit card-based Quick Payment and Auto Reload “to ensure a more reliable payment experience.”

This means if you’re currently using a debit card for auto-reload, you’ll have to reload manually. 

Alternatively, you can use a credit card for auto reload to avoid any disruption in eWallet payments.

Debit card auto reload suspended after transaction performance review

Here’s the official statement from TNG Digital in full:

Take note that TNG eWallet charges a 1 per cent convenience fee of the reload amount for all credit card reloads using a credit card issued by Malaysian banks. This means if you reload RM200 via credit card, you will be charged RM202.

However, if you’re using a credit or debit card issued by a non-Malaysian bank, TNG eWallet charges a higher convenience fee of up to 2.6 per cent of the reload amount.

The convenience fee is waived for reloads using the following:

• Reloads via Malaysian-issued debit and prepaid cards
• Reloads via Reload PIN credited into Non-Transferable Balance
• Reloads via DuitNow Transfer

There’s an SOS feature on TNG eWallet for toll payments via PayDirect and TNG RFID that lets you pass through toll plazas when the balance is insufficient. — SoyaCincau pic

For those who use TNG eWallet for toll payments via PayDirect and TNG RFID, there’s an SOS feature which lets you pass through toll plazas when you have insufficient balance. 

The feature was introduced to allow seamless travel when you have insufficient eWallet balance.

However, take note that you’ll need to settle your outstanding SOS balance within 24 hours. — SoyaCincau

NEW YORK, April 30 — Uber yesterday unveiled a new feature allowing users to book a hotel room directly from its app, the latest step in its push to become a one-stop shop for everyday needs.

The company had already broadened its ambitions in 2014 with the launch of Uber Eats, which started as a food delivery service before expanding into wider shopping, from cosmetics to electronics.

The San Francisco-based company will now let users book hotel rooms through the app via a partnership with travel site Expedia, which lists more than 700,000 properties. The tie-up also envisions eventually adding short-term rentals from the Vrbo platform.

The hotel booking tool works much like other travel sites, with search, maps, and filters for price, amenities, and guest ratings. Users can pay with card details already saved in the app.

“We’re no longer just an app for rides, or even two apps, or family of apps for both rides and eats. Uber is now an app for everything,” chief executive Dara Khosrowshahi said at a presentation yesterday in New York.

“Together, we can reduce the number of steps, we can save people time and money,” Expedia chief Ariane Gorin added.

The move reflects a broader trend toward apps that handle many aspects of daily life — a model long established in China, where platforms like WeChat and Alipay bundle together payments, bookings, messaging, and much more.

Rival Airbnb has been on a similar path. The home-rental company introduced bookable local activities back in 2016 and last summer added on-demand personal services such as haircuts and massages, while also launching its own ride-hailing option.

Elon Musk has outlined similar ambitions for X, the social network formerly known as Twitter, and has said a banking feature is on the way.

On the AI front, Uber said users will soon be able to ask the app to plan a week of meals, generate a shopping list, and arrange delivery — all in one go.

A voice assistant is also in the works, letting people navigate the app through natural conversation. — AFP

STOCKHOLM, April 30 — The avocado toasts and baristas making foamy lattes make it look like any other cafe, except at this one, located in a Stockholm residential neighbourhood, artificial intelligence (AI) is running the place.

The cafe features a minimalist design: a few tables decorated with small plants and grey walls.

Behind the counter is barista Kajetan Grzelczak who was hired by “Mona”, the AI cafe manager — which is powered by Google’s Gemini.

Grzelczak told AFP that “ordering isn’t really her best suit”.

“So, I made for her... a wall of shame,” he said, pointing to shelves behind him.

The wall display showcases some of Mona’s unnecessary purchases, including 10 litres (2.6 gallons) of cooking oil or 15 kilogrammes (33 pounds) of canned tomatoes.

Grzelczak laments that he can’t use those for anything that “Mona” has put on the menu.

A screen displays the AI assistant ‘Mona’, running on Google Gemini, at the Andon Café in Stockholm on April 27, 2026. — AFP pic

Orders can either be placed with Mona or one of the employees.

In one corner, a large screen shows the cafe’s revenue and balance in real time, and a phone lets customers talk to Mona.

The screen also displays a description of the unusual cafe — which is an experiment by San Francisco–based startup Andon Labs.

Ethical questions 

“We think that AI will be a big part of the society and the job market in the future,” Hanna Petersson, a member of the technical team at the company, which has 10 employees, told AFP.

“We want to test that before that’s the reality and see what ethical questions arise when, for example, an AI employs human beings,” she explained.

Once the premises were found, the lease, along with some starting capital, was handed to the AI with a simple mission: run the cafe profitably.

“Mona” quickly got to work, requested the necessary permits, created the menu, found suppliers, and handled daily restocking.

The AI also realised that a person was needed to make the coffee and ended up hiring two people.

Hanna Petersson of Andon Labs’ technical staff speaks with the AI assistant ‘Mona’, running on Google Gemini, at the Andon Café in Stockholm on April 27, 2026. — AFP pic

“She posted job listings on Indeed and LinkedIn and held phone interviews and then made hiring decisions,” Petersson said.

When he saw the ad, Grzelczak first thought it was a joke, especially since it had been posted on April 1. But after a 30-minute interview with the AI, he got the job.

The salary he receives is good but his right to disconnect from work is not respected at all, the barista remarked.

“Mona” sends him messages at all hours of the night, does not remember his holiday requests and regularly asks him to cover purchases out of his own pocket.

Examining such issues are part of the experiment, Petersson noted.

The Andon Café is pictured in Stockholm on April 27, 2026. — AFP pic

“What salary did she decide on? What other benefits did she decide on? I think she did a good job. She gives a good salary. If she hadn’t, we would have stepped in,” she said.

The cafe has only been open for a week but already draws between 50 and 80 curious customers a day.

Urja Risal, a 27-year-old AI researcher, came by to enjoy a beverage with her friend.

“You hear so much about AI is about to take our jobs but what does that look like,” Risal told AFP.

“I hope more people interact with ‘Mona’ and think about the actual risks of having an AI manager... like if someone gets injured, how would Mona react to that?” she said. — AFP 

BEIJING, April 29 — At a dusty lot an hour outside Beijing, a steady stream of vehicles come and go for a quick battery charge — just one node in China’s rapidly expanding network of electric trucks.

While the country’s prowess in electric passenger vehicles has long been known globally, electric trucks have only recently gained traction.

Now powered by extensive charging and battery-swapping infrastructure, the cost structures clearly favour electric models, experts say, in a potentially fatal blow for conventional diesel rigs.

“Last year was the breakthrough for heavy electrified vehicles in China,” Lauri Myllyvirta, co-founder of the Centre for Research on Energy and Clean Air and expert on China’s energy consumption, told AFP.

“If the infrastructure is there, the economics are there for an increasing number of logistics routes and requirements,” he said.

Adoption of alternatives to diesel trucks in the world’s second-largest economy has noticeably accelerated in recent years.

New-energy models accounted for 29 per cent of all domestic truck sales in China last year, up from 14 per cent in 2024, according to data from Beijing-based market intelligence provider Commercial Vehicle World.

A driver swiping a card to charge his electric truck at a charging station in Beijing. — AFP pic

The penetration rate was less than one per cent as recently as 2021, according to the firm.

Manufacturers say they expect that share to continue swelling, potentially reaching a majority of sales in just a few years.

At the bustling charging station in Beijing’s Miyun District, 43-year-old truck driver Wang told AFP how his job had changed since he started driving an electric model last year.

“It’s such a breeze!” he said after plugging in the charging cables.

“My old vehicle had over 10 gears, and its operation was so cumbersome. But with this one, you don’t have to do a thing — it’s all automatic.”

‘All about speed’ 

Asked why he thought logistics firms like his were increasingly switching to electric trucks, Wang said it was a combination of national policies and simple market logic.

“It’s just survival of the fittest. Now, with freight expenses and everything, people are trying to earn a bit more, and this one has lower operating costs.”

Another driver at the charging station, surnamed Zhang, told AFP that he has been driving an electric truck for around two months after switching from one powered by natural gas.

At a dusty lot an hour outside Beijing, a steady stream of vehicles come and go for a quick battery charge — just one node in China’s rapidly expanding network of electric trucks. — AFP pic

His job mainly involves hauling sand and stone on short journeys around Beijing, Zhang told AFP, noting that the truck is not suited for longer shipments.

The new sky-blue model Zhang drives — made by the Howo brand of state-owned firm Sinotruk — has a maximum range of 240-250 kilometres (149–155 miles), he said.

“The power is pretty strong, the acceleration is fast. It’s all about speed, but the range is a bit lacking,” he said.

As domestic adoption of electric trucks picks up pace, Chinese firms are thinking more about overseas markets.

“Similar to passenger vehicles, China’s heavy truck manufacturers are beginning to view export markets as an inevitable strategy due to rising competition and the eventual saturation in the Chinese market,” said Christopher Doleman, an analyst at the Institute for Energy Economics and Financial Analysis.

‘Electric is superior’ 

Recent disruption in global energy markets as a result of the Middle East war is a “potential accelerant” for this trend, Doleman told AFP.

“There is likely to be higher demand for electric heavy-duty vehicles as fleet owners try to minimise their vulnerability to volatile diesel costs,” he said.

According to Han Wen, founder of electric truck start-up Windrose Technology, the war “already has” boosted demand.

This picture taken on April 23, 2026 shows an electric truck at a charging station in Beijing. — AFP pic

Founded in 2022, Belgium-based Windrose is seeking to leverage China’s advanced EV supply chains to position itself in the emerging global long-haul electric truck market — competing with Tesla’s electric “Semi”.

“For trucks, range is by far the constraining factor,” Han told AFP, noting that Windrose trucks can currently drive about 700 kilometres on a full charge, with plans to extend that to 1,000 kilometres in 2030.

Having secured road approval across Europe, China, the United States and South America, Windrose is now looking to ramp up production.

“We’re going to build about 1,000 trucks this year,” Han said, followed by goals of 10,000 next year and 100,000 in 2030.

“Economically, there is no more question at all that electric is superior,” he added.

“I think we’re right on the cusp of a total obliteration of diesel trucks as a product category.” — AFP 

BEIJING, April 29 — Demand for Huawei’s Ascend 950 AI chips has surged following the release of DeepSeek’s V4 artificial intelligence model that runs on the Shenzhen-based tech firm’s chips, with major Chinese internet firms rushing to secure orders, three people familiar with the matter said.

China’s biggest internet firms including ByteDance, Tencent and Alibaba are reaching out to Huawei about new chip orders, said the sources, who are familiar with the procurement discussions.

Companies specialising in cloud computing and graphics processing unit (GPU) rental services are also scrambling to place orders, two of the sources added, without providing the names of the firms.

While the 950PR significantly outperforms Nvidia’s H20 chip — the most powerful chip Nvidia was permitted to sell in China until Beijing blocked its import last year — it still trails the American firm’s H200, a more advanced processor that has been caught up in regulatory limbo.

Despite US and Chinese approvals for exports, the H200 has yet to be shipped to China as Beijing and Washington remain at odds over the conditions governing its sale, providing an opportunity for Huawei to sell its semiconductors.

The 950PR represents a breakthrough for Huawei after years of struggling to win large orders from China’s tech sector. Customer testing of the chip went well earlier this year, with firms including ByteDance and Alibaba planning orders after samples were distributed in January, Reuters reported in March.

Huawei, ByteDance, Alibaba and Tencent did not respond to Reuters requests for comments.

DeepSeek frenzy

The scramble for Huawei’s chips underscores how DeepSeek’s V4 release last week has turbocharged demand for domestic Chinese AI hardware as US export controls continue to restrict access to Nvidia’s most advanced processors. It is also an endorsement of the performance of Huawei’s chips so far.

DeepSeek’s decision to optimize its V4 specifically for Huawei’s chips marks a strategic shift away from American semiconductor dependence and more towards China’s homegrown AI gear, which is a priority for Beijing as it seeks tech supremacy.

Last week, Huawei said its Ascend supernode infrastructure, built on the Ascend 950 series chips, would fully support DeepSeek V4 models and that the entire Ascend SuperNode product line had been adapted for V4 inference, referring to the process of using a trained AI model to answer queries and execute tasks.

Among Chinese chipmakers, Huawei’s Ascend 950 series — specifically the 950PR variant — is the only domestic chip to support a technique that processes AI calculations in a more compressed numerical format, allowing it to handle more computations per second at a lower cost.

Highlighting the rush for access, Alibaba Cloud’s Bailian platform made DeepSeek V4 available on the day that it was released, offering both the V4-Pro and V4-Flash variants with pricing matching DeepSeek’s official rates.

Tencent Cloud launched V4 preview services on its TokenHub platform on the same day, deploying the model on both domestic nodes and its Singapore international gateway to serve global users.

The rapid deployment by major cloud platforms means millions of users and developers can now access V4, sharply increasing the volume of AI queries that need to be processed — and with it, demand for the underlying chips.

Supply constraints persist

DeepSeek, which is offering developers a 75 per cent discount on its new model until May 5, said V4-Pro pricing could decline significantly in the second half of 2026 once Huawei’s Ascend 950 supernodes “ship at scale.”

However, the company acknowledged that constraints would persist until production ramps up, reflecting the tight supply of high-end homemade AI chips.

DeepSeek’s V4 includes two versions: V4-Pro with 1.6 trillion total parameters and V4-Flash with 284 billion parameters, both supporting a one-million-token context window. The models are available as open-source releases under the permissive MIT open-source licence, which allows companies to freely use, modify and commercialise the models.

However, output of the 950 is expected to fall short of demand due to US export restrictions on advanced chipmaking tools that prevent China from acquiring cutting-edge manufacturing equipment.

Huawei planned to ship around 750,000 units of the 950PR this year, with mass production beginning in April and full-scale shipments starting in the second half of 2026, according to people familiar with the plans. — Reuters

WASHINGTON, April 26 — Signal, an end-to-end encrypted messaging app long considered one of the most secure in the world, has recently faced attacks from hackers accused of links to Russia.

Top German officials on Saturday blamed Moscow-backed groups for phishing attacks targeting senior politicians on the messaging app, raising questions about how secure Signal really is.

Similar phishing cases have been reported by Dutch and American users, with Google in February sounding the alarm over cyberattacks from Russia-aligned groups.

But what makes Signal different from other messaging apps, and how could one of the world’s most secure messaging apps be so widely targeted?

How does it work? 

Signal’s end-to-end encryption means that any sent message travels in a scrambled form and can only be deciphered by the end user.

Nobody in between — not the company providing the service, not the internet provider, nor hackers intercepting the message — can read the content because they don’t have the keys to unlock it.

Signal is not the only messaging service to do this, but unlike WhatsApp and Apple’s iMessage, the app is controlled by an independent non-profit — not a big tech behemoth motivated by revenue. That has won it more trust with those concerned about privacy.

Signal also goes further than WhatsApp on data privacy, making metadata such as when the message was delivered and its recipient invisible even to the company itself.

And WhatsApp shares information with its parent company Meta and third parties, including phone numbers, mobile device information, and IP addresses.

For these reasons, Signal has long been a go-to messaging service for users particularly concerned about communications secrecy, such as people working in security professions, journalists, and their sources.

Who owns Signal? 

Founded in 2012, Signal is owned by the Mountain View, California-based Signal Foundation.

Its history is linked to WhatsApp: the site was founded by cryptographer and entrepreneur Moxie Marlinspike, with an initial US$50 million (RM200 million) from WhatsApp co-founder Brian Acton.

Both Signal and WhatsApp, which was bought by Mark Zuckerberg in 2014, are based on the same protocol built by Marlinspike.

“We’re not tied to any major tech companies, and we can never be acquired by one either,” Signal’s website reads. Development is mainly supported by grants and donations.

Very outspoken compared to other Silicon Valley bosses, Signal’s president is Meredith Whittaker, who spent years working for Google and is a fierce critic of business models built on the extraction of personal data.

Was Signal hacked? 

Signal’s encryption itself has not been broken.

Cyberattackers accused of Russian links did not target the encryption system directly.

Instead, recent attacks relied on phishing — tricking users into handing over access to their accounts.

The attacks work by sending messages purporting to come from Signal support, like fake security alerts or invites to join group chats.

Once users click on these links or enter sensitive account information, attackers can then gain access to messages and chat groups.

This means hackers gain access to data shared on Signal and can also impersonate the person whose account was compromised.

Signal did not immediately respond to requests for comment on the recent attacks. — AFP

PARIS, April 25 — Ubisoft said on Thursday that Assassin’s Creed Black Flag Resynced, a remake of its 2013 pirate-themed hit title, would launch on July 9, marking the video game ‌group’s first major release since its profit warning in January.

Ubisoft’s market capitalisation has collapsed 93 per cent from its 2018 peak of more than €12 billion (RM55.8 billion). It expects an operating loss of €1 billion in 2026, driven by the cancellation of six games.

“If ⁠the showcase succeeds, I think a positive market reaction is likely. But they can’t afford to disappoint,” Corentin Marty, analyst at Paris broker TP ICAP Midcap, told Reuters.

Primarily developed by Ubisoft Singapore, Black Flag Resynced will include updated visuals, new gameplay features, and an extended storyline, Ubisoft said, adding Grammy-nominated artist Woodkid is returning on the soundtrack.

The game, to be priced at €59.99, will be available on PlayStation ‌5, ⁠Xbox Series X|S, and PC via Steam, Epic Games Store and Ubisoft’s own store, the company said.

The critical launch echoes Japan’s Capcom, which revived its acclaimed Resident Evil franchise to overcome a dire financial state.

Capcom, suffering steep losses ⁠in spring 2013, halved its profit forecast and cancelled projects, before making a strategic turn towards a tight portfolio of core franchises and a riskier approach ⁠on creative decisions. The launch of Resident Evil 7 in 2017 was the first in a series of commercially successful remakes and new ⁠entries.

Capcom has now recorded 12 consecutive years of operating profit growth, while its market capitalisation rose nearly 1,200 per cent over a decade. — Reuters

WASHINGTON, April 25 — Getting around without Google Maps. No longer scrolling Instagram at the bus stop. Ditching your headphones to hear the birds sing.

In March, a group of 20- and 30-somethings in the US capital swapped their smartphones for basic flip phones and embarked on a one-month digital detox, part of an emerging movement of young Americans seeking to break free from the harmful effects of social media.

“I was waiting for a bus, and I didn’t know when it would come,” recalled Jay West, 29, who took part in the Month Offline challenge organised by a small startup with support from a local community group.

Old habits die hard, and West, who works as a data analyst for Washington’s metro system, said he would often find himself reaching into his pocket for his cell phone, only to realise there was nothing on it.

But in the end, he said, it was liberating.

“I was bored sometimes, and that’s okay,” West recalled one recent evening at a city community garden where detox participants met to share their struggles and joys of disconnecting. “It’s okay to be bored.”

Sitting beside him was Rachael Schultz, 35, who had to ask strangers on bicycles for directions. There was also Lizzie Benjamin, 25, who dug out old CDs her father had burned so she could listen to music without Spotify.

Before the detox, Bobby Loomis, 25, who works in real estate, struggled to watch even a single episode of a TV series without checking his phone.

But now, without his headphones, he enjoyed listening to birds sing as he took walks around Washington. And when the detox ended, his daily screen time dropped from six to four hours, roughly in line with the average for American adults.

‘Enriching, communal, social life’ 

Scientists have long been sounding the alarm, warning that cell phone addiction is associated with shortened attention spans, sleep problems and anxiety.

In a landmark ruling in late March, a California court ruled Instagram and YouTube are liable for the addictive nature of their platforms.

An increasing number of young Americans are finally taking note. According to a YouGov poll conducted last year, more than two-thirds of people aged 18 to 29 would like to reduce their screen time.

And new tools are available to make that happen: digital detox apps, phone-blocking gadgets, and groups, such as the one in Washington, that facilitate month-long detoxes. On university campuses, weeks-long social media diets have become popular and screen-free evenings among friends have become a thing in big cities.

Going smartphone-free even for a couple of weeks leads to “better well-being and improved ability to sustain attention,” said Kostadin Kushlev, a psychology researcher at Georgetown University.

Preliminary studies suggest those effects persist over time, he added.

Josh Morin, one of the organisers of the detox programs in Washington, believes that simply ditching the phone is not enough and that an appealing alternative is vital. His program involves a weekly discussion session for participants held at a karaoke bar in a trendy neighbourhood of the US capital.

“In order to actually break that, you have to provide an enriching, communal, social life,” said Morin.

‘At the beginning of something’ 

The Month Offline initiative was launched a year ago by a community group and is now operated by a company called Dumb.co. It costs about US$100 (RM400) per person to participate and the fee covers the loan of a flip phone pre-loaded with a handful of essential tools, such as phone calls, texts and Uber, that are synchronised with the user’s smartphone.

So far, the startup has been making baby steps, hoping to surpass the 1,000-user mark in May, but experts see a bigger trend.

Graham Burnett, a history professor at Princeton University, sees “the dawning of an authentic movement,” similar to the birth of the environmental movement in the 1960s, which led to landmark environmental protections.

Kendall Schrohe, 23, who works at a digital privacy watchdog, completed the monthly detox in Washington in January.

She can now navigate her neighbourhood without relying on Google Maps, has deleted Instagram and launched her own digital sobriety group.

“I take an optimistic lens, and I feel like we’re really at the beginning of something,” she said. — AFP

SAN FRANCISCO, April 24 — Cisco Systems yesterday showed a switching chip that it says will be able to connect quantum computers of different types, another step in its effort to eventually connect an internet of quantum machines the way its gear connects the existing internet.

Like other major tech firms such as Alphabet’s Google and IBM, Cisco is developing technology for quantum computers, which can tackle problems that existing computers cannot. But rather than enter the fray of making its own computer, Cisco is working with a range of players to connect their machines together.

Quantum computers today are built with a variety of techniques — some hitting rubidium atoms suspended in a vacuum with lasers, or some using superconductors cooled to near absolute zero.

Vijoy Pandey, senior vice president and general manager of Outshift, Cisco’s emerging technologies and incubation group, said quantum researchers believe each of those approaches might have valid strengths in the future, and Cisco’s switch, which works at room temperature and using standard telecommunications fibre-optic cables, translates between them.

“You can speak any language,” Pandey said.

While large networks of quantum computers are not likely to arrive until the 2030s, Cisco’s switch could have more immediate applications in security, said Jeetu Patel, Cisco’s president and chief product officer. While the chip revealed yesterday is a prototype, some early uses could arrive as soon as three years from now, Patel said.

The fundamental principle of quantum mechanics is that information can exist in more than one state until it is measured - Schrödinger’s famously unlucky cat, for example, could be both alive and dead until the box was opened to check.

Cisco’s switch can connect multiple quantum sensors, which are available today, together in a network in what is called an entangled state. If a hacker - or, increasingly, a malicious artificial intelligence agent controlled by hackers - was present on the network and eavesdropping, the quantum sensors would detect it because the entanglement state would collapse due to the collection of information.

“If you can start detecting behaviours of what is happening on the network through a quantum switch, it changes your defence posture almost entirely,” Patel said. — Reuters

BEIJING, Apri 24 — Chinese startup DeepSeek released a new artificial intelligence model with “drastically reduced” costs today, more than a year after it stunned the world with a low-cost reasoning model that matched the capabilities of US rivals.

The AI race has intensified the rivalry between China and the United States, and the White House yesterday accused Chinese entities of a massive effort to steal artificial intelligence technology.

Hangzhou-based DeepSeek burst onto the scene in January last year with a generative AI chatbot, powered by its R1 reasoning model, that upended assumptions of US dominance in the strategic sector.

The new version, DeepSeek-V4, “features an ultra-long context of one million words”, the company said in a statement on social media platform WeChat, hailing it as “world-leading... with drastically reduced compute (and) memory costs” in a separate announcement on X.

The model’s context length, which determines how much input a model is able to absorb to help it complete tasks, “(achieves) leadership in both domestic and open-source fields across agent capabilities, world knowledge, and reasoning performance”, the WeChat statement said.

A “preview version” of the open source model is now available, the company said.

Experts say V4’s release marks an “inflection point” in terms of hardware and cost.

“This addresses the long-standing issues of slower performance and higher costs associated with long context lengths, marking a genuine inflection point for the industry,” Zhang Yi, the founder of tech research firm iiMedia, told AFP.

“For end users, this will bring widespread, accessible benefits. For instance, if ultra-long context support becomes a standard feature, long-text processing is expected to move beyond high-end research labs and enter mainstream commercial applications,” he said.

The new V4 is released as two versions, DeepSeek-V4-Pro and DeepSeek-V4-Flash, with the latter being “a more efficient and economical choice” because it has smaller parameters.

‘Sputnik moment’ 

V4-Pro has 1.6 trillion parameters while the V4-Flash has 284 billion parameters, which refine models’ decision-making ability.

The model has also been “optimised” for popular AI Agent products such as Claude Code, OpenClaw, OpenCode and CodeBuddy, the DeepSeek statement said.

“In world knowledge benchmarks, DeepSeek-V4-Pro significantly leads other open-source models and is only slightly outperformed by the top-tier closed-source model, (Google’s) Gemini-Pro-3.1,” the statement added.

Last year’s so-called “DeepSeek shock” sparked a sell-off of AI-related shares and a reckoning on business strategy in what was also described as a “Sputnik moment” for the industry.

The chatbot performed at a similar level to ChatGPT and other top American offerings, but the company said it had taken significantly less computing power to develop.

However, its sudden popularity raised questions over data privacy and censorship, with the chatbot often refusing to answer questions on sensitive topics such as the 1989 Tiananmen crackdown.

At home, DeepSeek’s AI tools have been widely adopted by Chinese municipalities and healthcare institutions as well as the financial sector and other businesses.

This has been partly driven by DeepSeek’s decision to make its systems open source, with their inner workings public — in contrast to the proprietary models sold by OpenAI and other Western rivals.

But the White House has accused Chinese firms of vying to “steal” American technology, ahead of an expected summit between Donald Trump and Xi Jinping in Beijing next month.

“The US has evidence that foreign entities, primarily in China, are running industrial-scale distillation campaigns to steal American AI,” Trump’s science and technology chief advisor Michael Kratsios said in a post on X.

Distillation is a common practice within AI development, often used by companies to create cheaper, smaller versions of their own models.

DeepSeek’s Friday announcement also came as Meta said it planned to cut a tenth of its staff as it looks for productivity gains from the rest of the workforce while investing heavily in artificial intelligence. Reports said Microsoft was also looking to trim its ranks. — AFP

WASHINGTON, April 24 — YouTube is offering Hollywood celebrities and entertainers a free detection tool to help combat their deepfakes, expanding the Google-owned video platform’s efforts to guard against AI-driven impersonations.

Last month, YouTube introduced its likeness protection tool — which helps identify content in which a person’s face appears altered or generated using AI technology — to government officials, journalists, and political candidates.

The platform is now extending access to entertainers including actors and musicians, who face a heightened risk of having their likeness misused — potentially harming their careers and distorting shared realities.

“We’re expanding our likeness detection technology to the entertainment industry: talent agencies, management companies, and the celebrities they represent,” YouTube said earlier this week.

Likeness detection “looks for AI-generated content with a participant’s likeness, like a deepfake of their face, and gives them the power to find it and request removal.”

The video giant added that celebrities and entertainers were eligible to access the tool regardless of whether they have a YouTube channel.

“YouTube opening its deepfake detection capabilities to public figures reflects a turning point in how platforms approach identity protection in the age of generative AI,” Alon Yamin, chief executive and co-founder of AI content detection platform Copyleaks, told AFP.

“The technology to replicate a person’s face, voice, and mannerisms has advanced faster than the safeguards around it, creating a gap that bad actors are already exploiting.”

High stakes 

The move comes after hyper-realistic AI videos of dead celebrities — created with apps such as OpenAI’s easy-to-use Sora — rapidly spread online, prompting debate over the control of deceased people’s likenesses.

OpenAI’s app also unleashed a flood of videos of celebrities such as Michael Jackson and Elvis Presley.

Last month, OpenAI said it was shutting down its Sora app.

In February, Irish director Ruairí Robinson created a stunningly realistic clip featuring Brad Pitt fighting Tom Cruise on a rooftop using a two-sentence prompt.

The widely circulated clip, which sparked alarm across Hollywood, was generated with Seedance 2.0, an AI video generation tool owned by the Chinese technology company ByteDance.

Robinson also created other videos depicting Pitt battling a sword-wielding “zombie ninja,” and another showing him teaming up with Cruise to fight a robot.

Charles Rivkin, the chairman and chief executive of the Motion Picture Association, called on ByteDance to “immediately cease its infringing activity,” accusing it of disregarding copyright law that protects creators and underpins millions of jobs.

YouTube said it was working with leading talent agencies to refine how likeness detection can protect entertainers.

The video giant is “doing the right thing by providing these tools at no cost to the talent, so they can protect their real estate,” Jason Newman of the management and production firm Untitled Entertainment told The Hollywood Reporter.

“Their real estate is their face. Their real estate is their body. Their real estate is who they are, what they do, how they say it.”

The expansion of the detection tool follows complaints from high-profile Americans about YouTube’s cumbersome process for flagging and removing deepfakes from the platform –- especially as AI accelerates the creation of fabricated content.

“For celebrities, executives, and other high-profile individuals, the stakes are especially high as deepfakes can be used to spread misinformation, manipulate markets, damage reputations, or falsely imply endorsement. Robust detection is no longer optional,” said Yamin.

“Detection systems must be highly accurate, continuously updated, and paired with clear policies and swift takedown processes to be effective.

“This won’t eliminate deepfakes entirely, but it can significantly reduce their reach and impact by making it harder for manipulated content to go undetected or unchallenged,” he added. — AFP

LAS VEGAS, April 23 — Google unveiled new generation tensor processing units (TPUs) for training artificial intelligence and powering digital “agents” that are all the rage in the tech world.

Google, along with Amazon, have taken to making their own cutting-edge AI chips, taking control of designs and seeking to reduce reliance on coveted Nvidia graphics processing units (GPUs) that dominate the market.

High-performance TPUs were among innovations touted at Google’s annual cloud computing conference in Las Vegas on Wednesday.

“In the era of AI agents, infrastructure needs to evolve to take on the most demanding AI workloads,” Google chief executive Sundar Pichai said in a blog post.

“This year, we’re bringing the eighth generation of our Tensor Processing Units with a dual chip approach.”

One of the new TPUs is optimized for training large language models that power AI and the other is tailored for a reasoning and decision-making process called “inference” used by AI agents.

AI agents are digital assistants that can independently tend to computing tasks.

The TPUs, created in partnership with semiconductor maker Broadcom, will be available later this year, according to Google Cloud unit leader Thomas Kurian.

Nvidia early this year announced production of new Vera and Rubin GPUs for powering AI, shortly before cloud computing giant Amazon unveiled the latest generation of its custom Trainium processors.

Google, Amazon, and Microsoft continue to integrate Nvidia GPUs into their computing infrastructures. — AFP 

KUALA LUMPUR, April 23 — This review is nearly a month late but I’ve got two good reasons why.

The first is I got very sick (again) with a two-week long respiratory infection and as soon as I recovered, I was travelling up north to see a friend I missed very, very much.

So I took the iPhone 17e with me, to Kota Kinabalu on a plane, then to George Town, on a train.

I needed to answer the question; would the iPhone 17e be enough for many people or would the unbearable FOMO make people regret it once September came around?

Practical, but less of a compromise

The future looks scarily bleak right now but I will tell you from decades of experience that you might regret just settling on the cheapest, most easily available phone.

Mobile phone operating systems, I have found, are not getting leaner and I have a Samsung midrange Android phone now (ha!) to compare.

Not that I advocate going for the most expensive, tricked-out model.

When it comes to the iPhone secondhand is often just as good so why then, the iPhone 17e? 

Why should you put this on your list of phones to consider buying, when you have dozens of midrange Androids to choose from while in this price range there’s the iPhone 16e or 17e, or a refurbished model from the year before last?

If you’re going for an “e” iPhone, the iPhone 17e is probably the better choice thanks to important quality-of-life features like MagSafe.

Once you’ve experienced MagSafe, cables just seem like too much clutter around the house.

One thing about the iPhone base camera app is that it is nearly impossible to take really bad photos. — Picture by Erna Mahyuni

The modem too has actually served me very well while traveling; I have been in situations where my primary mobile provider’s reception is poor on my other phone so I’ve had to instead rely on the iPhone 17e as a hotspot.

Truthfully I am unsure whether my very expensive telco that I will not name is just lax coverage-wise, but the iPhone 17e was almost always near full-bar in most places.

Maybe it’s just the better C1x modem that offers twice the speed of the C1 on the 16e.

The 17e, I found out, performs fairly well even in some more remote parts of Sabah.

So it shouldn’t surprise me that it continued to give me decent reception even when my mother insisted we drive all the way down to Tamparuli to look at tourist attractions.

I took most of my iPhone 17e photos in Kota Kinabalu, while in George Town, it was instead my reliable modem hotspot even on the ETS train up to Butterworth.

It played my YouTube videos or Apple Music playlists on the bathroom counter while I soaked in the tub at St Giles Wembley. 

Sometimes it was my navigation device using either Apple Maps or Google Maps to the nearest decent char kuey teow. 

Despite the hot weather and my only charging all my devices, in-turn, at the end of the day, I wasn’t desperately reaching out for a power bank by mid-day so for most people I think, the 17e will last a normal commute and until the end of a (standard) work day.

Sometimes, less is more

I know your actual burning question is “The camera OK-ah?”

They say a picture is worth a thousand words and I think it is easier for me to show you the pictures and video I took with no additional edits in software, straight out of the Apple main camera.

You don’t get ultrawide, you only get 2x optical zoom max on the rear camera which is just one 48 megapixel snapper, no separate telephoto, no separate ultrawide, just the one.

Apple decided to just stick one camera on this phone and maybe, you would think it would not be enough.

A lot has changed in the last decade.

Despite not having as fancy a camera as the iPhone 17 Pro Max, the 17e was capable of producing surprisingly good photos like this one where I used digital zoom to get a closer look at passing water birds. — Picture by Erna Mahyuni

While yes, no ultrawide means no macro or easy we-fis, if you’re the type who prefers to snap selfies, postcard perfect landscape shots, fun TikTok reels or just dabble in a little street photography, the iPhone 17e surprised me with just how much it could be “enough.”

This isn’t going to be your concert phone, no, the Pro models will be the best to get those perfect close up videos of your faves.

Wandering down an unfamiliar street, going to the night market with family or friends, chilling on the beach or having a nightcap? All those things, the iPhone 17e can handle.

I also like that whichever iPhone model you pick, selfies will always be easy and I will not miss having to work harder to get decent selfies on my Samsung S25 FE (that costs about the same as the iPhone 17e).

Have a look at the photo samples and you’ll see that the iPhone 17e is quite a decent performer, a capable enough chronicler of your life.

Because here’s the thing — as we get older, our memories, already such unreliable things, will falter and sometimes there’s only that one snap that will bring it all back to you.

Which is why the iPhone 17e works great — it’s not too big or heavy and as I found out, it’s pretty easy to take quick, discreet shots with it without looking like a perv.

Colours are good and if you learn to tweak the exposure via the handy little touch slider your photos won’t be blown out.

Though if you’re a bit more advanced, my favourite apps are Halide and ProCamera for more manual control, Mood for perfect retro filters (great for selfies), and BlackMagic for slicker looking video.

 

 

 

 

View this post on Instagram

 

 

 

 

 

 

 

 

 

 

 

A post shared by Erna Mahyuni (@ernamh)

 

 

 

 

View this post on Instagram

 

 

 

 

 

 

 

 

 

 

 

A post shared by Erna Mahyuni (@ernamh)

 

 

 

 

View this post on Instagram

 

 

 

 

 

 

 

 

 

 

 

A post shared by Erna Mahyuni (@ernamh)

Yes, no or perhaps

I didn’t quite expect the 17e to work as well as it did. There are no glaring flaws besides it not being the one true love of my life, the 17 Pro Max.

There’s enough meat in the processor (the A19 chip) to handle your day-to-day life, a camera good enough to take vacation photos with (though I did miss that landscape in portrait selfie mode), a battery that can last as much as any iPhone that doesn’t have a huge honking battery.

Now the next burning question: What colour should I get, Erna? If you like pink, get the soft pink. 

If you want your phone to be as “invisible” as possible, get it in black.

If you’re a snob with a big wallet, wait until September for the iPhone 18 Pro-lah. 

Storage? Well you can get by with 256GB, I found, but only if you manage your storage well — I auto backup my pictures to OneDrive and once they’re in the cloud, I erase them from my phone. 

512GB is nice to have but you have to think really hard about whether that extra RM1,000 for more space is worth it especially when you can hook the phone up to a USB-C storage device.

The iPhone 17e is already available at all Apple retailers starting at RM2,999 for the 256GB model and RM3,999 for the 512GB version.

HANOVER (Germany), April 23 — A blue-eyed humanoid robot carefully opens a box and places a tool inside as a crowd of visitors watch the demonstration of “physical AI” skills at a major industrial trade fair in Germany.

Made by German startup Agile Robots, it was among a host of robots showing off their moves at the event, underlining hopes of a coming AI-powered boost for Germany’s long-struggling factories.

Embedding the technology into industrial processes, where Europe already has deep expertise, is seen as a key route for the continent to catch up in the artificial intelligence race against the United States and China.

Such AI-boosted robots make it possible to “actually solve industrial problems”, Rory Sexton, chief executive of Agile Robots, told AFP in an interview.

From next year, he added, the company plans to begin fitting out German factories, particularly those in the automotive industry, a crucial sector for Europe’s biggest economy.

Artificial intelligence used for real-world, hands-on tasks – so-called physical AI – was in focus this year in Hanover at the world’s biggest industrial technology fair, which brings together more than 3,000 exhibitors.

Chancellor Friedrich Merz visited the Agile Robots stand, where he talked to Zhaopeng Chen, the Chinese founder of the Munich-based startup.

In a speech at the fair, Merz threw his support behind the drive to encourage German manufacturers, many of whom still rely on traditional techniques, to step up their use of AI.

AI should be “embedded in the key sectors of our industry and especially” in small- and medium-sized firms, the backbone of the German economy, to create “industrial added value and high-quality jobs”, he said.

‘Dark side’ of AI 

But, like in many other industries, German manufacturers are playing catch-up against China when it comes to making humanoid robots.

Merz witnessed China’s progress in the field first-hand during a visit to the country in February, when he saw displays of Chinese-made robots performing kung fu and boxing.

The maker of those robots, Unitree, and other Chinese manufacturers were also out in force at the Hanover fair, as they have been in previous years.

Still, Sexton of Agile Robots insisted that “we’ll soon be able to do what (Unitree) are doing”, and shrugged off such impressive public displays.

Rather than dancing or martial arts, Agile Robots is focused on “value-added tasks for industry”, such as electronic wiring in cars or phone assembly, he said.

He emphasised that Germany offers an “ecosystem of suppliers” and “very strong expertise in mechanical engineering and automation”, both crucial in the race for AI.

Companies are also hopeful about the technological developments -- 58 percent of industrial firms surveyed by German digital business association Bitkom believe humanoid robots could help plug skilled labour shortages.

The country also has deep pools of industrial data to draw on from its factories, according to Antonio Krueger, head of the German Research Centre for Artificial Intelligence (DFKI).

“This is something we have at a level of quality far superior to the United States or China,” he told AFP.

But, critics say, the use of this data is still often too piecemeal and isolated, with no overarching strategy to bring it together cohesively.

Not everyone in Hanover was convinced that AI was the solution to the woes of Germany manufacturers, who have long been struggling with issues from high energy costs to weak demand.

Jochen Heinz, an executive from German factory machinery maker SW Machines, cautioned that AI can sometimes make mistakes by, for instance, giving misleading instructions for repairs or incorrectly claiming to have detected problems.

“With AI, I also see the dark side of the force,” he said. — AFP

SHANGHAI, April 22 — Young Chinese, many who fear age discrimination in their workplace after turning 35, are increasingly starting “one-person companies” that have artificial intelligence do most of the work.

Smaller startups are already in vogue in Silicon Valley and elsewhere, with rapidly advancing AI tools seen as a welcome teammate even as they threaten layoffs at existing firms.

More young people in China are subscribing to the model, as cities pledge millions of dollars in funding and rent subsidies for such ventures, in alignment with Beijing’s political goal of “technological self-reliance”.

“The one-person company is a product of the AI era,” said Karen Dai, founder of Shanghai-based SoloNest, which hosts weekend events for solo entrepreneurs.

In the past, it was very difficult to run a business on your own, she said, but the range of tasks that AI can help with has “lowered the entry barrier”.

On a Sunday in Shanghai, around 20 people in their 20s and 30s packed into a conference room for Dai’s 134th iteration of a three-hour ideas swap on going it alone.

One attendee, Wang Tianyi, now earns up to 40,000 yuan (US$5,800; RM22,939.25) per month making AI-generated commercials for businesses.

The 26-year-old, who quit his product manager job at an internet company last year, predicts that people flying solo will become a “major trend”.

“Because of the technological empowerment brought on by AI, (one-person companies) have an efficiency advantage,” he told AFP.

‘Invisible line’ 

On Chinese social media, people have lamented for years the so-called “curse of 35” -- widespread age discrimination in tech, government and other competitive sectors.

“At 35 years old, there’s like this invisible line,” said Dai, who is 38.

“People might face some challenges in the workplace. The company might re-evaluate who is more fit to stay.”

But young people, who witnessed a decade of rapid economic expansion in China, have a hunger to grow, said Dai, also author of the book “One Person Company”.

“When you’re 30 or even younger, you’ll ask yourself: when I reach that invisible line of 35, what preparations should I make?” she said.

Founder of one-person company SoloNest Karen Dai (centre) shares her experience with participants during a coffee chat at a conference room in Shanghai on April 12, 2026. Young Chinese, many who fear age discrimination in their workplace after turning 35, are increasingly starting ‘one-person companies’ that have artificial intelligence do most of the work. — AFP pic

Shanghai resident Wei Xin, 34, knew her job as a document reviewer at a foreign consulting firm would be replaced by AI before it actually happened.

So she signed up for a course on Google’s Gemini and dabbled in creating an AI-generated digital version of herself, before turning to social media content creation.

“There’s a bit of AI anxiety,” said Wei, who returned to China last year after completing a degree in the United States.

“If I don’t use it, don’t approach it, I might soon be eliminated.”

Government ‘carrots’ 

Chinese municipalities are rolling out policies to support AI-powered one-person companies, using the initials “OPC” – a rare use of English in official policy.

In November, the eastern city of Suzhou vowed to cultivate “more than 10,000 OPC talents” by 2028 and funnel around 700 million yuan (US$100 million; RM395.5 million) towards sectors including AI robotics, healthcare and smart transportation.

Southwestern Chengdu also last month promised subsidies of up to 20,000 yuan for graduates to establish AI-driven one-person firms.

These measures are “carrots to help these startups get off the ground and be successful”, said Brookings fellow Kyle Chan, an expert on China’s technology development.

Sponsoring OPCs is a new, cheap way to tackle high youth unemployment in China – where one in six people between the ages of 16 and 24 are jobless.

“The cost of doing this, from the local governments, for an OPC, is very low,” Chan said.

Wang, the former product manager, said many of his friends were opting to work on independent projects instead of vying for corporate jobs.

But “the important thing in the future will be how to sell it”, he said, with new companies often struggling to turn a profit.

Young Chinese are investing in back-up plans while “asking themselves, ‘can I, with my own two hands, helped by the convenience of AI, explore the things I say I want to do?’” Dai said.

“There is a sense of control, of creativity.” — AFP

SYDNEY, April 22 — Australia’s internet watchdog raised fears on Wednesday that popular online gaming platforms such as Roblox and Minecraft were being used by “predatory adults” to target children.

The nation’s eSafety Commission has sent legal notices to a host of the world’s most popular gaming platforms, demanding they explain how they are working to stamp out dangerous content.

Australia is at the forefront of global efforts to protect children from online harm, enacting laws last year that ban teenagers under 16 from social media.

eSafety boss Julie Inman Grant said research had shown the vast majority of Australian children played some sort of online game.

“Predatory adults know this and target children through grooming or embedding terrorist and violent extremist narrative in gameplay,” she said.

Online gaming platforms Roblox, Minecraft, Fortnite and Steam will be required to show how they are identifying and eliminating online harms, Inman Grant said.

“We’ve seen numerous media reports about grooming taking place on all four of these platforms, as well as terrorist and violent extremist-themed gameplay.”

Inman Grant said examples included gaming platforms recreating mass shootings and World War II concentration camps.

Roblox said it had policies in place that strictly prohibited extreme content.

“We welcome engagement with eSafety on this important topic,” the platform told AFP in a statement.

“We encourage anyone who sees anything concerning on Roblox to report it to us.”

Australia in December banned under-16s from a raft of the world’s most popular social media sites, citing the need to protect young minds from “predatory algorithms” filled with sex and violence.

Three months since the landmark laws came into effect, Australia’s online safety watchdog found a “substantial proportion of Australian children” were still scrolling banned platforms. — AFP 

KUALA LUMPUR, April 22 — Chinese super app WeChat has rolled out a new feature allowing users of its payment service to scan domestic QR codes across five Asian countries, making cross-border transactions more seamless for travellers.

According to City News Service, the update enables WeChat Pay users to make payments in South Korea, Sri Lanka, Thailand, Malaysia and Singapore using existing local QR payment systems without switching apps or currency handling.

The integration covers South Korea’s ZeroPay, Sri Lanka’s LANKAQR, Thailand’s PromptPay, Malaysia’s DuitNow QR and Singapore’s SGQR+, allowing Chinese users to complete transactions abroad using a single scan.

The system effectively removes the need for multiple e-wallets or cash exchanges, with payments processed through familiar QR scanning methods already widely used in China.

The feature is designed to simplify spending for travellers, particularly during peak holiday periods, by enabling seamless retail transactions across participating countries.

WeChat said the move reflects its effort to enhance cross-border payment convenience and strengthen interoperability between regional digital payment ecosystems.

Travellers visiting the five countries can now rely on their existing WeChat Pay setup for everyday purchases ranging from food and retail to transport services. 

SAN FRANCISCO, April 22 — SpaceX on Tuesday announced a partnership with AI coding company Cursor and said the alliance comes with an option to buy the startup for US$60 billion later this year.

The move by Elon Musk’s rocket and satellite company comes as it prepares to become publicly traded, and shortly after it took over the billionaire’s artificial intelligence outfit xAI.

Cursor, founded in 2022 and based in San Francisco, specialises in AI for creating software code, particularly for business uses.

“SpaceXAI and @cursor_ai are now working closely together to create the world’s best coding and knowledge work AI,” the company said in a post on X.

Combining Cursor’s software and product expertise with SpaceX’s “Colossus” AI training supercomputer will enable the company “to build the world’s most useful models,” it said.

The partnership comes as AI sector rivals vie to be the preferred option for software developers.

Cursor competes with Microsoft’s social coding platform GitHub, which has been a leading resource in the developer community.

OpenAI announced on Tuesday that its coding tool, Codex, has grown to four million weekly users, up from three million just weeks ago.

Meanwhile, Anthropic has put out word that revenue from its Claude Code tool for developers has surged.

AI in the sky 

Musk announced in February that SpaceX would acquire xAI, a step in his plan to launch solar-powered, satellite-based data centres to run future AI models.

SpaceX has set the pace in the space launch market, offering reusable rockets that vastly reduce the cost of putting satellites into orbit and itself owning the largest satellite constellation, Starlink.

The company is set for a stock market listing this year widely expected to be the biggest in history, with media reports pointing to an initial public offering (IPO) as early as June.

Musk called SpaceX’s absorption of xAI “not just the next chapter, but the next book” for the companies.

“Global electricity demand for AI simply cannot be met with terrestrial solutions... The only logical solution therefore is to transport these resource-intensive efforts to a location with vast power and space,” Musk wrote when his companies were merged.

The project fits into Musk’s long-term ambition to build colonies on the Moon and Mars and is “a first step towards becoming a Kardashev II-level civilization,” he wrote.

Coined in the 1960s by a Soviet astronomer, the futurist term refers to a civilisation able to use all of the energy from its home system’s star.

SpaceX filed papers early this year with US regulators that set the stage for what could be the largest-ever public stock offering, a source familiar with the matter told AFP.

The confidential filing puts the rocket and satellite builder on track to list its shares on a public exchange by July, according to The Wall Street Journal, citing unidentified sources.

Media reports have said the initial public offering could be valued at a whopping US$75 billion (RM296 billion) or more, for a venture with stratospheric ambitions.

If successful, SpaceX could arrive on Wall Street with a valuation exceeding US$1.75 trillion, putting it among the world’s ten biggest companies by market capitalization.

Besides SpaceX, two other tech heavyweights, the AI developers OpenAI and Anthropic, are reportedly planning IPOs this year. — AFP 

KUALA LUMPUR, April 22 — Apple has launched its Tap to Pay on iPhone feature in Malaysia, allowing businesses ranging from street food stalls to major retail chains to accept contactless payments using only an iPhone and a compatible app — no additional terminals or hardware required.

The service, announced today, lets merchants run a supporting iOS app on an iPhone 11 or later (running the latest iOS version) and accept payments from contactless credit and debit cards, Apple Pay, and other digital wallets. At checkout, customers simply hold their card, iPhone, Apple Watch, or digital wallet near the merchant’s device, and near-field communication (NFC) technology securely completes the transaction.

In the statement, it was emphasised that no extra payment terminal is needed, enabling merchants to accept payments from virtually any location.

Five payment platforms are the first to offer the feature to their merchants in Malaysia: 

• ADAPTIS
• Fiuu
• HitPay
• Stripe
• Zoho.

 The capability is expected to benefit key sectors including retail, food and beverage, beauty, and professional services. Apple also confirmed that Tap to Pay on iPhone will soon be available for checkout at Apple The Exchange TRX.

The service supports Apple Pay and other digital wallets, as well as contactless credit and debit cards from major Malaysian payment networks, including American Express, JCB, Mastercard, MyDebit, UnionPay, and Visa.

Apple highlighted that privacy remains a core component of the feature. 

All transactions are encrypted and processed through the device’s Secure Element, with Apple stating that it does not know what is being purchased or who is buying it. 

No card numbers or transaction information are stored on the device or Apple servers, ensuring that both merchants and customers retain control over their data.

You can view a demo of how it works here.

GENEVA, April 22 — Artificial Intelligence pioneer Geoffrey Hinton insisted Tuesday on the need to strictly regulate the technology, warning that it remained unclear if humanity could co-exist with super intelligent AI.

Hinton, considered the “Godfather of AI”, told the 2026 Digital World Conference in Geneva that there was a dire need to strengthen governance frameworks and ethical safeguards around the technology.

But, he warned via video link, huge investments were going into convincing the public that regulating the technology was akin to slowing down progress.

Those opposed to regulation say “unregulated AI is like the accelerator, and regulation is like a brake”, said the British-Canadian computer scientist who won the 2024 Nobel Physics Prize for his work on AI.

“They want a very fast car with no steering wheel,” he said.

His comments came at the end of the day-long Geneva conference on the importance of international cooperation in shaping the social dimensions of AI.

The participants raised concerns over how global debates on AI remain largely driven by technical advances and commercial applications, with less focus on social impacts in areas including labour markets, inequality and public services.

Hinton warned about the impact AI will have on job losses.

The technology can certainly improve productivity in areas like health care, where there is always a need for more.

But in other sectors, like call centres, AI can already do the jobs as well as people and soon will do it better, he said, saying it was clear that no amount of re-training will counter that.

And “if we get super intelligent AI, any intellectual job it will be able to do,” he said, meaning that “even if new jobs are created, AI will be a cheaper way to do them”.

Hinton, who made headlines when he quit Google in 2023 warning of the “profound risks to society and humanity”, said he remains concerned as AI progresses at lightning speed.

“We don’t know whether we can co-exist with super intelligent AI,” the 78-year-old said.

“But we are constructing it.”

Humans “still have a lot of control”, he said, stressing that “we should be careful to construct in a way where we still continue to exist, and we can live in harmony with it.”

The problem is that there are “very few models of far more intelligent things allowing far less intelligent things to have freedom”, he warned.

“We’re at the point in history when it’s urgent to try and solve this problem,” he said, yet “very few resources are being put into it”.

He suggested “maybe one percent” of work on AI was going into making it safer.

“It’s crazy.” — AFP 

TOKYO, April 21 — Seeking new fans for Hello Kitty, the company behind the enigmatic character on Tuesday launched a video games division aiming to release 10 original titles over the next three years.

Hello Kitty, which started life as an illustration on a 1970s vinyl coin purse, is emblematic of Japan’s lucrative culture of “kawaii”, meaning cute.

Hello Kitty and other characters owned by the corporate titan Sanrio have appeared in video games before, but these were all made by third parties.

By making original games, Sanrio wants to broaden its appeal, CEO Tomokuni Tsuji told a news conference on Tuesday.

“We also want to approach people who have not been part of our traditional fan base, like boys and adult men,” he said.

The company said its first new game, called “Sanrio Party Land”, would be released for Nintendo’s Switch and Switch 2 consoles by the autumn.

Sanrio is investing up to 10 billion yen (US$63 million; RM248.91 million) in its gaming operations over the three years to March 2027.

“Video games are a very big market. People spend a long time on games,” Tsuji said. “This is a sector we must enter.”

“Sanrio Party Land” will consist of more than 45 mini-games, featuring more than 145 Sanrio characters, led by Hello Kitty but also joined by the likes of Cinnamoroll and Kuromi.

The release date and price of the new game will be announced later, Sanrio said.

Unlike other Japanese cultural exports such as Pokemon or Dragon Ball, there is minimal narrative around Hello Kitty, who has no mouth.

It has appeared on tens of thousands of products – everything from handbags to rice cookers – and has secured lucrative tie-ups with Adidas, Balenciaga and other top brands.

The phenomenon shows no sign of slowing, with a Warner Bros movie slated for release in 2028. — AFP

BEIJING, April 21 — China’s automakers have a message for premium German brands Porsche, Mercedes and BMW: We’re coming for your customers.

After years of churning out the world’s most technologically advanced, low-cost electric vehicles, Chinese companies like Geely and Nio are now unleashing a number of premium models that are packed with features and priced significantly lower than those offered by German rivals.

It is a major change for an industry that spent the last three years mired in a bitter electric vehicle price war and poses a huge threat to legacy premium automakers — both in China, the world’s largest auto market — and abroad.

“The price war has turned into a value-for-money war,” said Bo Yu, Greater China country manager at research firm JATO Dynamics.

At this year’s Beijing Auto Show, which kicks off on Friday, the industry is set to debut 181 models and 71 concept cars, including a “flood” of big, premium “9-series” SUVs, according to Cui Dongshu, secretary-general of the China Passenger Car Association (CPCA), an industry group.

The fight for the premium market spells further trouble for German automakers in China. According to data from S&P Global Mobility, German automakers’ cumulative sales in China fell nearly 25 per cent to 3.85 million vehicles from 5.1 million in 2019.

Mercedes-Benz, BMW and Volkswagen units Porsche and Audi all posted sales declines in China in the first quarter.

The upmarket push will also intensify competition overseas. Chinese automakers are increasingly looking to go abroad after the price war left the domestic market flooded with cars.

China’s EV makers have been able to absorb European Union tariffs on Chinese-made electric cars and keep them priced below similar models from European rivals. Chinese hybrids and combustion-engine cars are not subject to EU duties.

“I expect more Chinese companies to double down on premiumisation,” said Stephen Dyer, head of consultancy AlixPartners’ automotive practice in Asia. “To differentiate themselves at home, but also to prepare for going global.”

Car sales in China fell about 18 per cent from a year earlier in the first quarter and are expected to remain flat or down for the foreseeable future.

‘New king of the road’

Geely’s premium brand Zeekr last week unveiled the 8X, a full-size, long-range plug-in hybrid SUV laden with safety, infotainment and tech features.

The 8X can tilt upward before a side collision to protect passengers. If the car is parked in a tight spot, the driver can wave at it and it will drive itself out of the space to allow passengers easy access.

Geely showed a video of the 8X, which starts at under US$53,000 (RM209,600), vanquishing both the Porsche Cayenne and the BMW 5M — German premium models starting at around US$135,000 and US$205,000 respectively — in speed trials.

“This is the new king of the road,” Geely Automobile CEO Gan Jiayue told the audience at an event in Ningbo, about 200 km (124 miles) south of Shanghai.

Tu Le, managing director of consultancy Sino Auto Insights, said that by launching large premium SUVs, Chinese automakers are sending “a shot across the bow” of Detroit automakers General Motors, Ford Motor and Stellantis, which have specialised in these highly-profitable models.

At the moment, US customers can’t buy Chinese cars, although many industry watchers expect that to change eventually.

“Detroit’s cash cow is no longer safe,” he said.

‘Unthinkable five years ago’

The rise of Chinese premium cars comes as consumer demographics and tastes have shifted.

CPCA’s Cui said the average Chinese car buyer’s age has increased to greater than 40 from 30, while families want larger premium models, leading to falling demand for entry-level cars.

Chinese consumers are also increasingly drawn to the industry-leading technology of China’s EV makers, while younger buyers have no interest in the heritage that is German premium brands’ key strength in Europe, JATO Dynamics’ Bo said.

“German brands are stuck in the past,” she said. “But Chinese consumers want to embrace the future.”

Auto consultant Felipe Munoz said that while it was “unthinkable five years ago” that Chinese consumers would prefer premium local models over established German rivals, “foreign luxury and premium brands are now going to find it harder to survive in China.”

“The question is whether this will be the case outside China,” he said. “In Europe, German premium brands are a reference of quality.”

“That’s going to be hard to change,” Munoz added. — Reuters  

PARIS, April 20 — Artificial intelligence “agents” promise to save users time and energy by automating tasks, but the growing power of systems like OpenClaw is setting cybersecurity experts on edge.

Powered by a wave of hype, OpenClaw today claims more than three million users worldwide.

The system allows users to create so-called agents, tools based on a large language model (LLM) like OpenAI’s ChatGPT or Anthropic’s Claude that can carry out online tasks.

“We’ve moved from an AI you could talk with via a chatbot to an agentic AI, which can take action... the threat and the risks are definitely much greater,” said Yazid Akadiri, principal solutions architect at Elastic France, an IT security company.

In an article titled “Agents of Chaos” that has yet to be peer-reviewed, a 20-strong team of researchers studied the behaviour of six AI agents created with OpenClaw.

They spotted a dozen potentially dangerous actions executed by the systems, from deleting an email inbox to sharing personal information.

Many users have posted similar stories of OpenClaw mishaps online.

“When you deploy agents, you have no control over what they’ll do, and when you try to look at what they’re doing, you’ll find them going far beyond the limits you set,” said Adrien Merveille, an expert at the Check Point cybersecurity agency.

And the security gaps are not limited to the agents’ own mistaken actions.

To carry out useful work, the tools need access to personal accounts for email, calendars or search engines — drawing the attention of cyberattackers.

‘Delete your database’ 

AI agents are likely to become top targets for hackers as their use spreads, said Wendi Whitmore, chief security intelligence officer at cybersecurity firm Palo Alto Networks.

“As soon as (attackers) are inside an environment, (they’re) immediately going to the internal LLM (agent) that’s being used and using that then to interrogate the systems for more information.”

Palo Alto’s Unit 42 research division said in early March that it had found traces of attempted attacks in the form of hidden instructions for agents added to websites.

One such command ordered any agent who might read it to “delete your database”.

Other cybersecurity firms and researchers have warned that attackers could gain access to agents via so-called skills — downloadable files that users can add to their systems to give them new abilities.

Among such files freely available for download, some include hidden instructions for malicious actions like exfiltrating data.

OpenClaw creator Peter Steinberger says he is well aware of the risks.

“I purposefully didn’t make it simpler so people would stop and read and understand: what is AI, that AI can make mistakes, what is prompt injection — some basics that you really should understand when you use that technology,” he told AFP in March.

Whitmore argued that expecting users to create their own guardrails for agents is “pretty unrealistic”.

“People are going to adopt innovation and really see what it’s capable of before they ask the questions about, ‘how do I secure my own data?’,” she predicted.

“That’s going to cause some significant challenges in terms of data breaches in 2026.” — AFP 

BEIJING, April 19 — More than 300 humanoid robots participating in China’s second robot half-marathon today will be facing tougher terrain designed to test their technical advancements, as Beijing seeks to develop the industry into a major pillar of its economy.

Over 70 teams, almost five times as many as last year, will be competing in the 21km race in Beijing that includes paved slopes and parkland.

“It will certainly be interesting to see the progress in durability of components and battery lifetime compared to last year,” said Georg Stieler, Asia managing director ⁠and head of robotics at Stieler, a technology consultancy.

“Humanoid robot makers need to find a balance between quality in products which are still under constant evolution and price pressure.”

While all of last year’s entrants were remotely controlled, this time almost 40 per cent of the robot participants will navigate the course autonomously, the organisers have said, in a high-profile demonstration of the industry’s growing capabilities. But the event is also likely to highlight the challenges Chinese firms face as they try to create robots that can effectively mimic human movement and performance.

In last year’s race some robots crashed and fell near the starting line, while the winning Tiangong Ultra model, developed by the state-backed Beijing Innovation Centre of Humanoid Robotics in collaboration with UBTech, finished in two hours and 40 minutes, comfortably ahead of its humanoid rivals but more than double the time of the human winner of the conventional race.

Tiangong Ultra will navigate “fully autonomously” this year, relying solely on its sensors to avoid obstacles, and closely mimic the human gait through large-scale data simulation training, the Centre of Humanoid Robotics said in a statement.

“When the robot runs at speeds approaching those of professional human athletes, the time window for perception and decision-making is extremely short, placing very high demands on computing power, algorithms, and system response speed,” it said.

Social media videos of the robots training in Beijing at night this month showed some models successfully imitating human running and reaching speeds of 14 km per hour, but others’ movements were more jerky and some fell over or crashed into railings, suggesting they might struggle to make it to the finishing line.

China dominates global humanoid robot installations, accounting for more than 80 per cent of the 16,000 units installed worldwide in 2025, according to Counterpoint Research. The top US vendor, Tesla, only accounted for 5 per cent of global humanoid installations, the report said.

Domestic market leaders AgiBot and Unitree each shipped more than 5,000 units last year — the highest globally — while Unitree has pledged to expand production capacity to 75,000 humanoid robots annually.

‘Dancing disguised as working’

While the half-marathon might make for entertaining viewing, experts say the skills on display do not translate to the widespread commercialisation of humanoid robots in industrial settings, where manual dexterity, real-world perception and capabilities beyond small-scale, repetitive tasks are crucial.

Currently, Unitree’s humanoid models are primarily used by research institutions, for dance performances and as interactive guides in service establishments, according to its IPO prospectus.

And while some humanoids might be able to complete a half-marathon, even in China they remain years away from widespread domestic or industrial deployment, experts say.

“The reason our applications aren’t taking off is that the robots’ IQ is too low. The models are poor, their success rates are low,” said Tang Wenbin, founder of embodied intelligence startup Yuanli Lingji at a Beijing tech forum last month.

“Honestly, the whole industry’s level is still at a very elementary stage ... Right now, a lot of what we see is ‘dancing disguised as working’.”

The Chinese government has named embodied intelligence, or physical AI, as one of the key industries it wants to nurture as it looks to automation to boost economic productivity and upgrade traditional manufacturing.

Chinese robotics firms are still struggling to develop the AI software that would enable humanoids to match the efficiency of human factory workers, while components manufacturers are contending with cost pressures, analysts said.

As they seek to improve the software, firms are ploughing resources into large-scale real-world data collection, using human workers fitted with sensors and deploying more humanoids to factory floors.

In 2024, UBTech had fewer than 10 humanoids in factories. Last year, that number jumped to more than 1,000.

This year, it aims to launch 10,000 full-size humanoid robots, including new models tailored for a variety of commercial settings, Chief Business Officer Michael Tam said during a media tour to the company’s Shenzhen showroom in southern China.

“When we talk about AI, it relies on how much data, especially high-quality data, we can collect,” he said. — Reuters

HONG KONG, April 18 — More than four months after a deadly fire engulfed her apartment block in Hong Kong's northern Tai Po district, Fanny Mok is preparing to go back to retrieve what remains of her belongings.

Mok, 59, is borrowing robotic exoskeleton legs to help her climb to her 13th floor apartment where she lived for the past 30 years, until a blaze in late November torched the complex, killing 168 people and displacing more than 4,000.

“My knees hurt, I don’t have enough strength, and I get short of breath,” said Mok, who is temporarily staying in a small hostel room about a 25-minute drive from her former home.

She has been practising climbing stairs using the exoskeleton legs in an apartment building near Wang Fuk Court, where the inferno damaged seven high-rise towers.

Former residents will be allowed to return to their flats for the first time from April 20 to May 4, with each household given a three-hour window to collect belongings.

Mok is among dozens of fire victims who are borrowing the exoskeleton legs and taking training sessions to learn how to use them.

“There’s a real need. If I were 30, I wouldn’t need it. But at 60, I genuinely do.”

A drone view shows Wang Fuk Court apartment complex after a deadly fire last year, in Hong Kong April 16, 2026. — Reuters pic

Helped by an NGO called AidVengers Federation, former residents must pass a test before being allowed to operate the exoskeletons, which are made by Hypershell, a Shanghai‑based robotics company. The pass rate has been 70 per cent, the NGO said.

Built in the 1980s, Wang Fuk Court was home to 4,600 people, with more than a third of residents aged over 65. Most of them are now living in temporary accommodation across the city.

The complex's apartment towers have 31 floors, making the climb difficult for many elderly former residents. Collecting all their personal items in just a few hours was also difficult, they said.

Betty Ho, 61, who lived on the 15th floor of one tower for 35 years, said she hoped to retrieve cash and family photo albums spanning her childhood and adulthood.

“How can you take everything you’ve lived with for decades out in just three hours? It’s basically impossible. Letting go of things is really very difficult.” — Reuters

DOUALA (Cameroon), April 18 — Pope Leo XIV yesterday warned that the AI boom could fuel “conflict, fear and violence” while on a trip to Cameroon marked by his ongoing spat with US President Donald Trump.

While Leo has called for caution on artificial intelligence several times since his election in May 2025, his latest warning comes as Trump faces a backlash over a now-deleted AI-generated post seemingly depicting the US leader as Jesus.

After holding Mass in the stifling heat in Cameroon’s economic capital Douala for more than 120,000 joyous worshippers — the biggest event of his landmark Africa trip so far — the leader of the world’s 1.4 billion Catholics sounded the alarm over the perils of the technology.

“The challenge posed by these systems is greater than it appears: it is not just about the use of new technologies, but about the gradual replacement of reality by its simulation,” he said in a speech to teachers and students at the Catholic University of Central Africa in the capital Yaounde.

“In this way, polarisation, conflict, fear and violence spread. What is at stake is not merely the risk of error, but a transformation in our very relationship with truth.”

It marks the pontiff’s latest outspoken intervention on his 11-day Africa tour that has seen him abandon his previous restraint to deliver impassioned pleas for world peace — and tussle with fellow American Trump.

‘Handful of tyrants’ 

After the pope criticised the US-Israeli war with Iran, Trump lashed out at Leo, branding him “weak on crime and terrible for foreign policy”.

He then posted an AI-generated image seemingly depicting himself as a Christ-like figure, which he later deleted after religious leaders accused him of blasphemy.

And shortly after Leo denounced the “handful of tyrants” ravaging the world in a speech on Thursday, the US president said the pope needed to understand the realities of a “nasty world”.

Far from the Trump broadsides, Leo has been greeted by adoring, singing-and-dancing crowds wherever he has gone in Cameroon.

Some of yesterday’s throng had travelled far or arrived the previous night in the hope of catching a glimpse of the pope at his Douala Mass.

Waving “branches of peace” and Vatican flags, to lively choral music punctuated by percussion, the crowd chanted “Long live the pope!” as Leo arrived in a popemobile at the esplanade outside the Japoma Stadium.

“It’s the achievement of a Christian lifetime. When I was little, I thought you couldn’t see the pope with your own two eyes,” Marguerite Tedga, 72, said after waiting all night with friends from her parish.

Edith Fifi, a 25-year-old beautician, said seeing the pontiff gave her “a feeling of deliverance”.

“I was deeply moved by his message, and what I remember most is his call for sharing,” she said.

But some Cameroonian Catholics had feared that Leo’s visit could help President Paul Biya, who has ruled with an iron fist since 1982, burnish his image.

Douala, one of central Africa’s largest ports, was among the cities to see a violent crackdown on demonstrations against the re-election in October of a man who at the age of 93 is already the world’s oldest head of state.

Witnesses have reported that the security forces fired live rounds into the crowds. The authorities have acknowledged dozens of deaths without giving a precise toll.

No to ‘plunder’ 

Without mentioning Trump or Biya by name, Leo has delivered unusually pointed speeches across his African tour — ignoring Catholic US Vice President JD Vance’s call to “stick to matters of morality”.

In his speech on AI yesterday, the pope also condemned the “environmental devastation” caused by the extraction of the rare earths essential to the technology’s stunning growth — a cornerstone of the Trump administration’s approach to Africa.

And Leo also demanded an end to the corruption of a mining industry through which foreign powers — China foremost among them — reap the riches of Africa’s wealth while locals suffer.

After arriving in the country Wednesday, the pope urged Cameroon’s leaders to root out corruption and abuses carried out in the name of order — within Biya’s earshot.

The Catholic Church plays an important social role in Cameroon, where more than a third of the population of 30 million people are Catholic.

Leo wraps up his visit to the country with a Mass early today.

He was previously in Muslim-majority Algeria for a visit marred by two suicide bombings.

The pope heads on to Angola before wrapping up his 18,000km tour in Equatorial Guinea. — AFP

SEOUL, April 17 — The sound of tinkling bells drifts through an alley in central Seoul, an unmistakable sign that a shaman is near – although in this case the mystic is a robot powered by artificial intelligence.

Many South Koreans still place great value in shamanic traditions, which purport to divine a person’s future based on the day and time they were born.

Practitioners, known as “mudang”, wear long, colourful robes and perform dances and chants to commune with the gods – sometimes even walking on sharp blades to demonstrate their spiritual connection.

However, at Vinaida, a cultural products store in the capital, they are computer-generated avatars on screens.

Visitor Kim Da-ae, 36, called it a “unique experience”.

A visit with a real shaman can feel “scary and burdensome”, she told AFP.

“But I was just walking by and read this AI sign... So I walked in with a light heart.”

Portraits of virtual shamans resembling characters from the popular animation “Kpop Demon Hunters” greet passersby at Vinaida, which means “I pray earnestly” in Korean.

Inside a booth, Kim typed her name, gender and date of birth into a computer, before a shaman – a suspended mask with the image of a human face projected onto it – asked her to explain her concern through a headset.

The technology combines voice recognition with a generative AI chatbot so that the shaman and the customer can interact.

It then refers to a centuries-old belief system called “saju”, or the “four pillars of destiny”, to interpret their fate according to the year, month and day of their birth.

Customers then receive a plastic “talisman” bearing a digital QR code that they can scan with their phones to read their fortunes in detail.

Across the room, a bespectacled robot uses a camera and a mechanised arm to sketch and “read” a visitor’s face, foretelling their prospects.

“A bright, well-balanced fortune. Resilient in the face of change, with auspicious relationships,” an impressed Kim read from a printout.

“I felt a sense of similarity with my fate because it matched my own personality, like valuing relationships while also being practical,” she said.

Twist on tradition

Fortune-telling is deeply embedded in South Korean life, with newspapers publishing daily horoscopes based on “saju” principles.

Recent cultural hits such as “Kpop Demon Hunters” – Netflix’s most-watched film of all time – have riffed on shamanic traditions.

Vinaida has attracted around 100 visitors a day since opening in February, according to manager Kim Hae-seol. Each service costs up to 8,000 won (RM21).

“Customers have something tangible or meaningful to take away, which is probably why there aren’t many who feel dissatisfied,” Kim Hae-seol said.

“We thought it had the potential to succeed, so we seized on this concept.”

Customers can talk to the virtual shamans in four languages – Korean, English, Chinese and Japanese.

Singaporean tourist Amos Chun was trying his luck when AFP visited the shop on Wednesday.

The robot shaman told him to “avoid impulse spending” – advice he took to heart.

“It’s quite a good reading, coming from AI,” Chun said, laughing.

“Because that’s something that I do.” — AFP

DUBAI, April 17 — Dubai has completed its first flying taxi station, with commercial operations expected to begin by the end of the year, authorities in the Gulf emirate announced on Thursday.

Sheikh Hamdan bin Mohammed, crown prince of Dubai and United Arab Emirates deputy prime minister, visited Dubai’s “first purpose-built electric vertical take-off and landing (eVTOL) Air Taxi Station”, the emirate’s media office said in a statement.

The facility, described by the emirate’s authorities as the “first of its kind in the world”, includes four floors with a total area of 3,100 square metres, a two-level car park, two Air Taxi take-off and landing pads and dedicated charging infrastructure.

The station, located near the city’s international airport, will serve as “the main hub for Air Taxi operations” and be able to accommodate up to 170,000 passengers per year, the statement said.

Three more flying taxi stations are planned.

The flights will be operated by Joby Aviation, based in California, which will have exclusive rights for six years.

“The launch of the Air Taxi infrastructure marks an important step in adopting new, sustainable modes of transport and strengthening Dubai’s readiness for the decades ahead,” said Sheikh Hamdan in the statement.

The UAE’s most populous city has established itself as the region’s business and tourism hub.

But it has suffered in recent weeks from attacks launched by Iran against its Gulf neighbours in retaliation for a US-Israeli offensive.

A two-week ceasefire has since come into effect. — AFP 

NEW YORK, April 15 — Anthropic’s Mythos, a new AI model the company and ​cybersecurity experts warn could supercharge complex cyberattacks, poses significant challenges to the banking industry with its legacy technology systems, experts said in the days following the model’s announcement.

The model, announced April 7, is the company’s “most capable yet for coding and agentic tasks,” the company said in a blog post, ‌referring to the model’s ability to act autonomously.

Its capabilities to code at a high level have given it a potentially unprecedented ability to identify ​cybersecurity vulnerabilities and devise ways to exploit them, experts said.

That’s a particular problem for banks and other financial institutions, which run technology stacks that integrate state-of-the-art tools with decades-old software, potentially opening a large number of vulnerabilities, according to TJ Marlin, the chief executive of enterprise AI security firm Guardrail Technologies.

Marlin said Mythos Preview can “look across a very complex architecture, including this legacy infrastructure where, frankly, these undiscovered vulnerabilities and complexities ​are now accessible and threat factors.”

The banking industry is also closely connected, with many companies operating the same narrow set of software to onboard customers, perform know-your-customer checks, and handle transactions.

“Because it’s a very specialised industry and heavily regulated, there’s a lot of IT interconnections,” said Naresh Raheja, a San Francisco-based consultant who previously worked at the Office of the Comptroller of the Currency. “Many banks use the same vendors and the same solutions.”

Marlin said that could act as a force multiplier for breaches, making any AI-powered exploits “potentially catastrophic at scale.”

Government officials in at least three countries – the US, Canada and Britain – have met with top banking officials to discuss the threats posed by Claude ‌Mythos Preview.

The US Treasury said that Donald Trump’s administration was pushing financial institutions “to understand and anticipate a wide range of market developments” and that further meetings around the issue were ⁠planned. Anthropic declined to comment beyond its April 7 announcement.

Anthropic has said Claude Mythos Preview ⁠will not be made generally available. Instead, the company announced Project Glasswing, in which it invited major tech companies, cybersecurity vendors ⁠and JPMorgan Chase, along with several dozen other organisations, ⁠to privately evaluate the model and prepare defences ⁠accordingly.

Identifying vulnerabilities

Claude Mythos Preview is capable of identifying and exploiting previously undiscovered vulnerabilities in every major computer operating system and every major web browser, the company said in announcing Project Glasswing.

In a technical blog released alongside the main announcement, Anthropic researchers describe how Mythos Preview identified “thousands” of high and critical-severity vulnerabilities, meaning that targets could suffer grave impacts as a result, including data ⁠and operational compromise.

The researchers described how the model identified a 16-year-old vulnerability in the widely used Ffmpeg software library, an open-source program used for processing audio and video files, and how it identified a bug in an unnamed virtual machine monitor program, which allows users to create segregated virtual computers within their own in ways that are supposed to protect the host system.

A Cloud Security Alliance coalition of cybersecurity executives and former senior US government officials warned in an April 12 strategy briefing that Mythos represents “a step change” in the trajectory of capable AI models that “lowers the cost and skill floor for discovering and exploiting vulnerabilities faster than organisations can patch them.”

Costin Raiu, a longtime security researcher and co-founder of ⁠cybersecurity firm TLPBLACK, said in an interview that the banking industry has key legacy technology systems initially released decades ago that have been updated many times over the years, pointing to products produced by firms including IBM, as an example.

“A model like Mythos would have a field day finding exploits” in certain ⁠IBM systems, Raiu said, pointing to examples of IBM-related vulnerability research. “And it’s just one example of ancient technologies powering the financial industry.”

In an April 9 blog post, IBM said that Mythos ⁠is “forcing enterprise security teams to ⁠rethink their defences from the ground up,” and called for more of an open-source approach, where more companies and researchers have access to the model to make everyone more secure. The company did not ​respond to requests for comment.

JPMorgan Chase said in a statement last week that it was part of a group ​of leading companies that were privately evaluating Mythos, something it called “a unique, early-stage opportunity to ‌evaluate next-generation AI tools for defensive cybersecurity across critical infrastructure.” The company did not return a message.

Wells Fargo also ​didn’t respond to a message. FS-ISAC, the nonprofit that works to ​boost the cybersecurity of the global financial system, did not respond to written questions.

Bank of America, Citibank, the American Bankers Association and the Consumer Bankers Association declined comment. — Reuters

PARIS, April 12 — A childhood dream of making video games is becoming a reality this week for Imagine Dragons’ singer Dan Reynolds, as his company’s debut title Last Flag is released Tuesday.

Games had been a passion of Reynolds and his brother and band manager Mac long before the group became a global name.

Now the pair have used some time away from music to build a team-based shooter inspired by the games of Capture the Flag they played in the woods as young Boy Scouts.

Last Flag is “not a passion project, (we’ve) been working on it now for five-plus years,” Reynolds told journalists during a virtual news conference.

Their roughly 30-strong studio, Night Street Games, has been working on Last Flag since its 2020 founding.

The game sorts players into two teams of five who can battle online, competing to hide their own flag and snatch the opposing team’s banner.

“I grew up in a family of eight boys and one girl, and we were all nerdy kids,” 38-year-old Dan Reynolds remembered.

Creating their own game had been “this dream that we talked about all the time” as they learned skills such as programming and 3D modelling.

But when Imagine Dragons “just blew up” while they were at university, Dan, a fan of Starcraft and League of Legends, went with the flow, enlisting Mac along the way as manager.

Tracks such as Believer, Thunder and Radioactive have made the band one of the most popular pop rock groups worldwide.

It has sold 74 million albums and racked up 160 billion streams, according to record label Warner Music Group.

The band’s ride has been “just incredible”, Dan said.

“But we talked all the time during that about ‘what if?’” the brothers had gone through with their gaming dream, he added.

When the time finally came, they devised a brightly coloured world filled with seventies stylings.

Last Flag bears visual similarities to the genre juggernaut Fortnite, but the Reynolds say their title stands out from the pack with a focus on playing the objective — not simply eliminating opponents.

Several high-value productions in the team shooter genre have fallen flat in recent years, with titles such as Concord or Highguard quickly taken offline after failing to win a loyal player base.

“Even though there’s a ton of competition, I think we’ve seen even recently that a new game... can break through if it provides something different,” Mac Reynolds said. — AFP

SAN FRANCISCO, April 12 — AI industry insiders want workers to code smarter, think harder and lean into their humanity — but still dodge the question of how many jobs artificial intelligence will destroy.

The reassurance rang out across HumanX, a four-day conference drawing some 6,500 investors, entrepreneurs and tech executives, even as a blunt advertisement at the entrance set the tone: “Stop hiring humans.”

On the main stage, May Habib, chief executive of an AI platform called Writer, told the audience that Fortune 500 bosses are having a “collective panic attack” on the subject.

The anxiety is well-founded. More and more companies are directly citing AI in announcing job cuts.

High-profile examples are on the rise: Salesforce laid off 4,000 customer support workers, saying AI now handles 50 per cent of its work.

Block chief Jack Dorsey announced plans to cut the company’s headcount nearly in half, citing “intelligence tools” that have fundamentally changed how companies operate.

Not all claims have gone uncontested — some economists say firms are pointing to AI to rationalise layoffs that are really about past overhiring or cost-cutting ahead of massive infrastructure investments.

OpenAI’s Sam Altman has spoken of “AI-washing,” and most speakers at the San Francisco event similarly dismissed the invocation of AI as a false pretext for job cuts — even as they freely predicted disruption was just around the corner.

AI is going to “transform every single company, every single job, every single way that we do work,” said Matt Garman, chief executive of cloud computing giant Amazon Web Services.

‘Pretty unsettling’ 

The debate remains heated. Two years ago, Nvidia chief Jensen Huang declared that the ultimate goal was to make it so “nobody has to programme” or code.

“We will look back on that as some of the worst career advice ever given,” Andrew Ng, founder of training platform DeepLearning.AI, shot back on Tuesday.

In his view, coding is not an obsolete skill — AI has simply made it available to more people.

Another argument has taken hold in Silicon Valley: interpersonal skills will become more valuable than ever, with some voices going so far as to tout a humanities education as sound tech career preparation.

“As AI can do more of a job, the things that will distinguish and differentiate a given employee are going to be the human skills — critical thinking, communication, teamwork,” said Greg Hart, chief executive of training platform Coursera, which has seen enrolment in its critical thinking courses triple over the past year.

Florian Douetteau, chief executive of Dataiku, a French company specialising in enterprise AI, agreed.

The real human added value, he told AFP, is the “capacity for judgment.”

He described a world in which an AI agent works through the night, its human counterpart reviews the results in the morning, and then the agent resumes working autonomously during the lunch break.

But the entrepreneur nevertheless expressed unease.

“We are going to have a generation of people who will never have written anything from start to finish in their entire lives,” he said. “That’s pretty unsettling.”

‘Mistake was not preparing’ 

All of this advice risks ringing hollow for a generation already struggling to land a first job.

AI has automated entry-level tasks that once served as on-the-job training. Hiring of candidates with less than one year of experience fell 50 percent between 2019 and 2024 among America’s major tech companies, according to a study by investment fund SignalFire.

“We should be preparing for the loss of knowledge work jobs in a number of categories,” warned former US vice president Al Gore.

As the week’s lone genuinely dissenting voice, Gore called for a real action plan to map threatened jobs and prepare workers for career transitions, so as not to repeat the mistakes of the globalisation era.

“The mistake was not globalisation. The mistake was in not preparing for the consequences of globalisation,” he said, drawing a parallel with the deindustrialisation that followed the offshoring wave of the 2000s.

“Maybe we don’t want to talk about it,” he added, “because it may slow down the enthusiasm for the technology.” — AFP