Tsinghua's AI healthcare system test bridges hope and caution amid global scrutiny

For an 80-year-old Beijing resident surnamed Ye, a hospital visit starts long before she reaches the door. At home, she logs into a mobile platform, types in her symptoms, uploads past prescriptions and test reports, and receives a preliminary assessment.

"It organizes my medical history and even flags issues I might have overlooked," she said. "I feel more prepared and less anxious about the visit."

The tool is part of Agent Hospital, an AI-driven healthcare support system developed by Tsinghua University's Institute for AI Industry Research. The program has been undergoing extended trials in six Beijing hospitals and two in the Guangxi Zhuang autonomous region. Developers hope it will make visits easier for patients while helping doctors with diagnostics and follow-up care.

The latest phase of testing opened the system to the general public, widening participation beyond earlier trials that focused on medical professionals and university staff. Researchers are learning how it performs among people with different levels of education, income, and familiarity with technology.

"We want to know whether it can function in community clinics as well as major hospitals," said Liu Yang, the institute's executive dean and the project's lead scientist. He added that medical services should be widely available.

For patients like Ye, one of the most useful features has been continued guidance after an appointment. Hospitals rarely provide that kind of support once patients walk out, she said, while the digital platform gives her a sense of reassurance.

Ma Weizhi, an assistant professor at the institute and a core team member, said the system can compile patient histories, generate records automatically, and point out possible risks. It includes portals for patients, physicians, and hospital managers, allowing each group to handle appointments, decision-making tools, or resource coordination.

Despite appreciating the system's efficiency, Ye is mindful of its broader implications. She wonders who would be liable if an AI recommendation caused harm. She also worries about how her personal data might be used and who can access it.

These concerns echo the caution voiced by global health experts. Peiling Yap, chief scientist at the Geneva nonprofit HealthAI, said artificial intelligence offers significant potential but also raises issues of accountability and privacy. Without clear oversight, she warned, systems may prove unsafe or ineffective.

In response to such challenges, project lead Liu Yang said that "the system strictly follows national regulations regarding data management." Those requirements were reinforced in a guideline released in November by the National Health Commission, which called for tighter security, stronger protection of personal information through the establishment of a clinical data authorization system, and a negative list for data management.

Yap said China's growing experience with regulation could benefit other countries if lessons are exchanged more broadly.

International interest has followed. During a visit to Tsinghua AIR on Nov 24, Italian Minister of Health Orazio Schillaci praised the project, saying it could work with existing healthcare systems to improve efficiency and assist with elderly care and physician training. He expressed hope for deeper Sino-Italian cooperation in digital health.

Mo Kai, a senior consultant at the Singapore-based regional association APACMed, described AI medicine as an area where China has notable advantages and said such projects are likely to play an expanding role in partnerships with developing nations.

Liu views Agent Hospital as groundwork for larger change. He said the initiative aims to link medical and engineering expertise and support the country's push toward a more digital healthcare system.