At a time when China is accelerating efforts to integrate artificial intelligence into healthcare, one of the country's most prominent medical experts is drawing a clear line.

"I refuse to introduce artificial intelligence into hospital medical record systems," Zhang Wenhong, a leading infectious disease expert and professor at Fudan University, said at a public forum in Hong Kong on Jan 10. His concern, he said, is what could be lost when machines begin to shape clinical judgment.

Doctor Zhang is one of China's most prominent infectious disease experts. He rose to national and international fame during the COVID-19 pandemic due to his role as a public-facing medical authority. He became a viral sensation early in 2020 for his blunt, "no-nonsense" communication style. He famously ordered that senior doctors and Communist Party of China members — rather than just junior staff — must take the frontline shifts in isolation wards.

Zhang's remarks sparked widespread discussion across China's medical and technology communities. They came as governments and investors pour resources into "AI+healthcare", and as tech leaders make increasingly bold claims about the technology's potential.

"Doctors must possess clinical judgment abilities that surpass those of AI," Zhang said.

While he did not deny the technology's usefulness — noting that he has used it himself — Zhang insisted that human expertise must remain central. "I can tell where AI is wrong," he said. "I won't be misled by it."

His primary concern is the impact on medical training. If AI-generated outputs become default "standard answers" in diagnosis and treatment, young doctors may never develop their own capacity for critical evaluation. "Without systematic training, doctors will lose the ability to judge whether AI's conclusions are correct," he said.

Zhang said he believes future physicians must master the skills of assessing the reliability of AI diagnoses and managing complex cases that algorithms cannot handle.

"That is the professional foundation behind technological convenience," he added.

Yet in hospitals across China, AI is already reshaping clinical practice.

At the Affiliated People's Hospital of Ningbo University in Zhejiang province, an oncology tool called PANDA analyzes plain CT scans to identify pancreatic cancer risk.

According to the hospital, the system has reviewed more than 180,000 abdominal and chest scans, flagging more than 20 cancer cases that initially went unnoticed.

"These patients came in with symptoms like bloating or nausea and did not undergo specialized pancreatic exams," said Zhu Kelei, director of the hospital's hepatobiliary and pancreatic surgery department. "In several cases, the CT images did not trigger concern until the AI system flagged them. I can say that AI entirely saved their lives in these cases," he said.

Expectations are also high at primary-level hospitals, where physician shortages and heavy patient loads are chronic problems. At Chuiyangliu Hospital affiliated with Tsinghua University in Beijing, a locally developed system known as Agent Hospital is undergoing real-world testing.

"The daily pressure here is immense," said an internal medicine doctor involved in the trial." Patients struggle to get appointments. AI assistance could help me see more people because it reduces the time I need to identify patients' conditions during consultations."

The AI system is a digital platform that allows doctors to track patients' medical histories, quickly locate suitable treatments and receive clinical suggestions.

For patients, the same system can provide guidance that may simplify examination procedures and help them describe symptoms more accurately when needed.

Hu Guodong, deputy secretary of the China Center for Information Industry Development, said AI healthcare has become a global focus for both capital and regulators. The global market, he said, was valued at under $30 billion in 2024 and is estimated to reach $500 billion by 2033.

In China, central and local governments are rolling out "AI+" initiatives, with ministries and municipal authorities launching pilot programs and funding schemes. "More than 100 AI medical device projects are now underway nationwide," Hu said.

Wu You, an associate professor at Tsinghua University's medical school, said AI implementation in the country typically begins in hospital departments with standardized data, high workloads and relatively low risks.

"At the national level, policy directives have clearly identified medical imaging, electrocardiography, clinical laboratory testing and pathological diagnosis as priority areas," Wu said.

"In practice, AI has been most widely adopted in medical imaging and rapidly expanding digital pathology. It is also advancing healthcare quality management and patient-facing services to improve operational efficiency," she said.

"AI in Chinese healthcare is usually rolled out from pilot programs in top-tier hospitals to regional medical consortiums, rather than being deployed directly in primary care settings such as community health centers and rural clinics," she added.

Still, skepticism remains strong within the medical community.

"I do not agree that AI will replace doctors in the short term," said Gao Wen, director of technology transfer at Capital Medical University. "Not every medical problem requires AI. Some technologies appear advanced but offer limited real benefit to healthcare."

As AI systems move deeper into diagnosis and treatment, regulatory challenges are becoming increasingly difficult to ignore.