Trading ‘Dr Google’ for ‘Dr AI’: Why more Malaysians are self-diagnosing using chatbots — and what medical experts fear most

KUALA LUMPUR, July 14 — As artificial intelligence (AI) tools rapidly  integrate into daily life, a troubling trend is emerging: an increasing number of people are turning to algorithms to self-diagnose ailments and self-prescribe medications.

This shift from “Dr Google” to “Dr AI” raises urgent questions about patient safety and the future of clinical care.

The scale of the trend is evident in recent data; a study from the Journal of Health, Wellness and Community Research in Pakistan found that between 44 per cent and 59 per cent of the general population trust AI over trained medical professionals to guide their health decisions.

In Malaysia, the reliance is even more pronounced. The Telenor Asia Digital Lives Decoded 2025 survey revealed that 67 per cent of the general population — and a staggering 89 per cent of internet users — actively rely on AI tools.

Approximately 81 per cent of these users use AI primarily for information and answers, often using traditional search engines merely to verify AI-generated responses.

To explore the implications of this trend, Malay Mail consulted medical experts on whether this reliance endangers lives by erasing the critical human touch essential to healthcare.

The peril of the ‘digital diagnosis’

Consultant public health physician and Perak State Health Department director, Dr Feisul Idzwan Mustapha, warns that while AI is an excellent informational resource, it is fundamentally not a doctor.

“The biggest risk is not AI replacing doctors, but people placing too much trust in AI-generated advice,” he cautioned, noting that such reliance often leads to misdiagnosis, inappropriate self-medication, or fatal delays in seeking professional treatment.

Prof Datuk Sri Dr Alwi Mohamed Yunus, a senior consultant cardiothoracic surgeon at the National Heart Institute, echoed these concerns.

He said that AI often overlooks the nuanced, individual health factors a professional considers during a physical examination.

“Doctors consider many factors, including the patient’s medical history, risk profile, medication use, laboratory results and imaging findings,” Dr Alwi explained.

“Often, important information is overlooked by the patient and only surfaces on specific questioning.”

Datuk Dr Thirunavukarasu Rajoo, president of the Malaysian Medical Association (MMA), argued that the danger lies in the confusion between information and diagnosis.

“The greatest risk is not necessarily that AI provides completely incorrect information. Rather, people may mistake information for a diagnosis,” he said.

He stressed that AI cannot assess vital signs, perform physical exams, or take legal and ethical accountability for treatment decisions.

“Medicine is more than matching symptoms to possible conditions. It requires clinical assessment, experience, judgment and accountability.”

Defining the boundaries of AI use

The experts agree that AI has a place in healthcare, provided it is used for general education rather than clinical decision-making.

Dr Feisul acknowledged that AI can be a useful starting point for health literacy, but warned: “AI should never be the sole basis for diagnosis, treatment decisions, or stopping prescribed medications.”

Dr Alwi added that medical advice must be hyper-personalised, particularly for patients with comorbidities like diabetes or kidney disease.

He warned that AI-generated reassurance can be dangerous, especially in time-sensitive emergencies.

“In heart disease, delay can be dangerous because conditions such as heart attack, severe valve disease or advanced heart failure can be time-sensitive,” he said.

Dr Thirunavukarasu suggested that AI is best used to help patients better understand their conditions and prepare more meaningful questions for their doctors.

However, he cautioned that the quality of an AI response depends entirely on the user's input, which is often incomplete or inaccurate.

The regulation debate: Innovation versus safety

When asked if these tools should be regulated, the experts offered differing perspectives.

Dr Feisul and Dr Thirunavukarasu argued for a focus on “responsible use” and digital literacy over strict restriction.

They suggested that AI platforms should be transparent about their limitations and clearly signpost users toward professional care.

“Public education is equally important,” Dr Thirunavukarasu said.

“People need to understand that AI can provide health information, but it cannot replace medical examination, investigations or clinical judgment.”

Dr Alwi, however, advocated for formal regulation, particularly when AI influences diagnosis and treatment.

He pointed to the need for national guidelines that mandate clinical validation and continuous monitoring to ensure these tools remain safe and accurate for the local population.

Harnessing the tool, not the master

Despite the risks, there is a consensus that AI, when used as an assistant, can revolutionise healthcare.

Dr Thirunavukarasu explained that AI is already proving its worth in analysing medical images, detecting diabetic eye disease, and improving clinical documentation.

Dr Alwi believes AI can deliver safer, faster, and more personalised care, provided the human remains the final authority.

“The principle must always be clear: AI-assisted, not AI-replaced,” he said.

“The final clinical decision should remain with qualified healthcare professionals who understand the patient’s full condition and can apply clinical judgment, empathy and accountability.”

Still, the goal is a synergy where technology handles the data, and humans handle the care.

As Dr Feisul put it: “The future is not AI versus doctors; it is AI working alongside healthcare professionals to deliver better care.”