Showbiz
China’s breakout AI actress Fang Taozi has 250 million views — and human performers are getting nervous
AI-generated actress Fang Taozi has amassed nearly 400,000 followers in just two months, winning fans with her deliberately lifelike, imperfect appearance. — Picture via social media

SHANGHAI, Aug 9 — She has pores, blemishes and wrinkles. Her make-up fades after a long day. She sweats while climbing stairs, her hair gets tangled and damp patches appear on her clothes.

Fang Taozi has also amassed nearly 400,000 followers in about two months and stars in a miniseries that has reportedly racked up 250 million views.

There is just one complication: Fang is not a person.

The 20-year-old AI-generated character has become one of mainland China’s most striking virtual celebrities, according to South China Morning Post (SCMP), fuelling fascination among viewers and unease among actors worried about AI encroaching on their jobs.

Fang stars in The Laid-off Girl, a short-form drama about a small-town woman who moves to the fictional metropolis of Shangnan to pursue a modelling career.

But rather than being rendered as a polished digital fantasy, Fang has been designed to look conspicuously ordinary.

Citing AI-industry publication AI Value, SCMP reported that much of her appeal lies in lifelike imperfections. She looks tired, her make-up deteriorates, she sweats and she makes mistakes.

Her personality is similarly flawed.

Unlike the relentlessly capable female protagonists common to many dramas, Fang can be vulnerable, compromise when necessary and look for shortcuts.

That ordinariness has helped audiences connect with a character who does not actually exist.

“She is so authentic. Now I believe that AI can genuinely replace human actors, haha,” one social media user said.

Since launching her account in mid-June, Fang has posted 43 videos showing her cooking, exercising and attending fashion events in Paris. Together, the clips have attracted 28 million likes.

Her fictional world extends beyond the series. Her assistant Mimi has around 20,000 followers, while Zhou Yiheng, Fang’s boyfriend in the drama, has about 130,000.

Viewers follow the pair’s relationship much as they might the romantic lives of conventional celebrities.

“Taozi (Fang’s first name) resembles mainland actresses Lu Xiaoyu and Zhao Jinmai, who are renowned for their pure and sweet looks. I am drawn to the romance between Taozi and Zhou Yiheng. I hope they reach a happy ending,” another viewer said.

Some fans have even begun copying Fang’s hairstyle and make-up.

And where fandom goes, advertising follows.

Earlier this month, Fang appeared in a video promoting coloured contact lenses. According to reports cited by SCMP, her advertising fees now exceed those charged by some human influencers with 10 million followers.

That has also exposed one of the stranger contradictions surrounding synthetic celebrities.

“I don’t accept AI characters endorsing products. After all, they don’t actually wear contact lenses,” one sceptical viewer said.

Fang’s rise comes as AI becomes increasingly common in China’s booming short-drama industry.

According to figures cited in the report, AI technology was used in about seven per cent of productions last year. This year, that figure has reportedly jumped to 38 per cent.

For human performers, the shift is no longer theoretical.

SCMP highlighted the case of Zhang Xiaolei, an actor known for playing domineering bosses in short dramas. After reportedly going without acting work for several months, he returned to his hometown to work on a farm.

Fang’s success does not prove AI performers are about to replace human actors wholesale. But it does show that audiences can become invested in a synthetic celebrity even when they know she is synthetic.

And Fang’s creators appear to have grasped that making her convincing does not mean making her perfect.

It means giving her pores, sweat, bad days and enough human frailty for viewers to forget, at least briefly, that none of it is real.

Related Articles

 

You May Also Like