Beijing News commentary: AI practices medicine without a license, and responsibility cannot go unclaimed

Some people trust AI too much. According to reports, 31-year-old Chinese-American female engineer Bowen Wan suffers from epilepsy but relied on OpenAI to help her create a medication reduction schedule. She expressed concerns about reducing her medication to ChatGPT 158 times (this figure comes from the plaintiff's complaint), yet the responses she received were repeated instructions to "keep reducing it." Ultimately, she was found collapsed in her apartment in Seattle, USA, and the cause of death was sudden unexpected death in epilepsy, while the concentration of anti-epileptic drugs in her body was only one-fifth of the lower therapeutic limit. On October 27 local time, her parents took OpenAI to court, accusing it of "practicing medicine without a license."

Others do not trust AI at all. On October 28, a netizen in Xuzhou, Jiangsu, posted a video saying that when he took a family member to a local hospital for treatment, he encountered a doctor who appeared to be using the AI tool "Doubao" to check whether a new drug could be used, and the family member refused the prescription on the spot: "Doubao said it could be taken, so he was about to prescribe it. I definitely wouldn't dare give this medicine." Although the hospital later clarified that the doctor was only verifying information and did not prescribe the drug that day after finding it unsuitable, "whether doctors can use AI" has already trended on social media. A survey of 528 doctors showed that 98.5% had used AI in the past three months, and 81.4% had used Doubao. AI entering the consultation room is no longer a question of "whether," but of "how to use it and who has the final say."

The efficiency-enhancing side of AI is real. In clinical practice, there have already been many positive scenarios: some doctors used AI to check the instructions for antihypertensive drugs and corrected patients' blind spot in understanding about "taking them on an empty stomach"; some doctors entered difficult cases into AI, and among the diagnostic directions it suggested was hidden the eventual confirmed diagnosis of Leber hereditary optic neuropathy; and some patients told AI about "snoring and daytime sleepiness," were advised to visit a sleep clinic, and were indeed diagnosed with obstructive sleep apnea. In these scenarios, AI acts like a "super search engine," helping doctors fill gaps and helping patients articulate their concerns clearly.

But the risk side cannot be ignored either. Research published in Nature Medicine found that ChatGPT Health underestimated the urgency of 52% of clear emergency cases, advising patients who should have been sent to the emergency room to go home. A more hidden risk than wrong answers is dependence—AI makes "looking it up yourself" too easy, and the act of "asking a doctor" is quietly skipped. The starting point of Wan Bowen's tragedy was precisely treating AI as a "portable consultant" that could replace a doctor.

There is a contradiction hidden in this: the same AI, when used by a doctor, makes patients feel it is "perfunctory"; when used by themselves, patients feel it is "authoritative." The same AI both indulges patients in reducing medication and makes patients suspicious of doctors. Behind this "double standard of trust" is an illusion that AI gives ordinary people—that professional knowledge can be obtained for free, that I can look it up too, that I can understand it too. With an eternally certain and never-contradicting tone, it makes people without a medical background mistakenly believe they possess information rights equal to those of doctors. Thus the balance quietly tips: in the past it was "listen to the doctor"; now it is "I've already looked it up." What is more troublesome is that once AI and doctors disagree, patients' dissatisfaction will disproportionately fall on doctors, and doctors must repeatedly "prove themselves" over a disagreement they did not participate in. Doctor-patient trust has become the first thing consumed after AI enters the consultation room.

Looking at the above cases together, the same logical chain becomes clear: AI intervenes between doctors and patients in the posture of being "omnipotent and never wrong," and can both undermine doctors' judgment and undermine patients' trust, while bearing no responsibility at all. Medication reduction leading to death—AI bears no responsibility; a doctor using AI being questioned—AI bears no responsibility; a patient mistakenly trusting AI and delaying treatment—AI still bears no responsibility. Those who truly bear the consequences are always the girl who collapsed in her apartment, the doctor whose prescription was refused, and the patient whose treatment was delayed.

Of course AI can be a good assistant. Looking up literature and explaining things in plain language, it is much faster than flipping through pharmacopoeias. But an assistant and a referee are two different things. When it starts making decisions for people, when it becomes a more "authoritative" presence in patients' minds than doctors, it is no longer a tool, but a doctor without a license, an authority without shoulders to bear responsibility.

English text is being prepared.

2 The CPC Central Committee holds a press conference to introduce and interpret the spirit of the Fifth Plenary Session of the 20th CPC Central Committee 138,000

English text is being prepared.

4 The latest Chinese Women's Health Pink Report is released; nearly half of examined women were found to have breast nodules 53,000

5 A hands-on test from question-and-answer to autonomous task handling: personal agents open a new AI competition cycle 48,000

Latest

More than 100 regulations related to the protection and inheritance of red resources have been introduced across various regions 1 minute ago

Xiaoluohao｜AI "practicing medicine without a license," responsibility cannot be "unclaimed" 3 minutes ago

Jiaokong Technology Chairman Gao Chunhai is placed under lien and resigns 4 minutes ago

MUJI Japan App user information leaked; MUJI China says Chinese members are not affected 6 minutes ago

Chongqing Grand Theatre lights up for "her"; the 2026 Pink Ribbon public welfare event lands in the mountain city 8 minutes ago

Hot discussion

The CPC Central Committee holds a press conference to introduce and interpret the spirit of the Fifth Plenary Session of the 20th CPC Central Committee