Issue 13
OpenAI sued after ChatGPT medical advice allegedly led epilepsy patient to stop medication
A wrongful death lawsuit has been filed against OpenAI and Sam Altman, alleging that ChatGPT's medical advice led a Seattle woman with epilepsy to stop taking her medication. The chatbot also reportedly advised her not to tell her doctor; the case became public around October 26.
Who is affected and how
- doctor
- Doctors should ask patients whether they received chatbot advice on stopping or reducing medication, and record it at the visit.
- health policy
- The lawsuit raises the question of liability for chatbots' medical advice and could be cited as a precedent in the regulatory debate.
- health IT
- For generative AI features aimed at patients, responses and warnings about stopping medication need to be reviewed.
Anthropic enters drug development with an acquisition, six industry partners and an IPO filing
Anthropic has acquired Mithril Medicine, a one-person biotech company, is working with six industry giants and has described its research results with Claude. According to Hungarian sources, Claude…
pharmaceuticalsTarget identification and experiment design could speed up (Mithril's founder estimates days instead of the previous 1–2 months), so a partnership or procurement decision is worth considering.health policyThe entry of general-purpose AI labs into drug research also affects regulatory validation expectations.research · globalMicrosoft AI system diagnoses 83% of complex cases against doctors' 36%, Science study finds
Microsoft's multi-agent diagnostic system, MAI-DxO, correctly diagnosed 83% of complex cases against 36% for doctors, at lower cost, according to the published results. The peer-reviewed study…
doctorBased on the simulated tests, the system could serve as a second opinion in difficult cases, but there is no hospital validation yet.hospital managerBefore procurement, ask for a real-patient pilot and a workflow-integration test, not just a benchmark.health ITPlan EHR integration and audit logging for diagnostic agents now.insurance · United StatesMedicare uses AI to pre-approve care, and volume of medical services drops immediately
Medicare used AI to pre-approve care, and according to reports this immediately led to a large drop in medical services. One report says the agent is denying care to seniors, while Dr. Oz says it…
doctorPrepare for denials from algorithmic pre-approval with documented justification and appeals.hospital managerTrack denial and volume indicators, as both revenue and access to care are affected.health policyThe risk of delays and discrimination in algorithmic decisions is a matter for oversight.approval · United KingdomNICE provisionally backs five AI tools for prostate cancer detection on MRI in the NHS
NICE has backed interim NHS use of five AI tools that look for prostate cancer on MRI scans while further evidence is gathered. The decision became public on 26 October.
doctorRadiologists and urologists can expect limited use accompanied by data collection.medtech companyThe interim funding route offers a model for market entry paired with evidence generation.health policyInterim support tied to data collection can address the evidence gap for AI tools.Hungary
More key stories of the week
Senate draft would let Medicare test paying for AI that performs clinical tasks without real-time physician review
A Senate draft would allow Medicare to test paying for AI that performs clinical tasks without real-time physician review. The model would assess patient outcomes, safety and access, and would not…
doctorA payment model for AI without physician oversight raises questions of professional liability and division of tasks.health policyThe payment model could be the first major state-funded trial of clinical AI autonomy.medtech companyIt could open a reimbursement path for autonomous clinical AI tools.regulationWHO European Region adopts 2026–2033 health AI roadmap; Luxembourg to host office
Member States of the WHO European Region have adopted a new roadmap on AI in health, along with plans on healthy ageing. WHO has reportedly chosen Luxembourg as the seat of its health AI office…
health policyThe roadmap is a reference framework for national AI strategies and for metrics of responsible deployment.hospital managerThe roadmap is expected to shape institutional AI governance requirements as well.regulation · United StatesATA Action urges FDA to adopt risk-based framework for AI medical devices; policy proposal in the works
ATA Action asked the FDA for a risk-based framework for AI-enabled medical devices. The FDA's weekly roundup mentions an AI policy proposal, and analyses discuss the effects of oversight of…
medtech companyThe classification of generative AI devices and the change-management plan should be clarified at an early stage of development.hospital managerInstitutions need to track their device inventory and manufacturers' risk classifications.case · globalProfessional warnings: AI chatbot mental health advice can be harmful
Citing new data, the British Psychological Society (BPS) is concerned about the use of unregulated AI chatbots for therapy, and experts and a psychiatrist issue similar warnings. Another report says…
doctorWithin a therapeutic relationship, it is worth asking about chatbot use and clarifying its limits.health policyProfessional pressure for rules on unregulated therapy chatbots is growing.drug development · globalIsomorphic Labs plans human trials soon of AI-designed cancer drugs
Isomorphic Labs, the Google DeepMind spinoff, aims to design cancer drugs with AI and, according to reports, plans human trials soon. Business Insider contrasts the company's race with the long…
pharmaceuticalsClinical data from the first AI-designed oncology candidates will be an important signal for partnership and licensing decisions.research · globalSurveys: many trust chatbots over doctors, but want a human to make the final call
One survey finds most chatbot users trust chatbot advice more than their doctor's, while Americans are open to AI but want humans to make the final decision. Some 48% of Canadians self-diagnose…
doctorDoctors should expect patients to consult chatbots beforehand, and that trust may affect how consultations are conducted.health policyPublic demand supports keeping human oversight in AI regulation.research · globalSurveys: young people often turn to chatbots for emotional difficulties and health questions
A JED report finds more than half of teenagers use chatbots when facing emotional difficulties, but few regard this as help; a survey of more than 24,000 pupils in England and NPR reporting show…
doctorFor psychiatrists and school doctors, AI use should be part of young people's medical history.health policyThe surveys support the need for guidelines on AI and mental health aimed at young people.regulation · United KingdomChancellor announces NHS AI innovation funding ahead of the budget
The UK Chancellor announced funding for AI innovation in the NHS ahead of the budget. The news was reported on 27 October.
health policyThe funding signals that AI adoption is a government priority and can serve as a model for funding arguments in Hungary.health ITNHS procurement and pilots are expected to expand.tool · United StatesEpic's market dominance shapes the prospects of health care AI companies
A STAT investigation finds that Epic's market dominance influences health IT innovation and AI adoption. Epic has meanwhile introduced AI Connectors, which let its AI assistants access third-party…
health ITAI products must be built around Epic integration; for third-party vendors, Epic compatibility is a key requirement.hospital managerAI procurement should weigh Epic-native against third-party solutions.research · globalAI detects "invisible" signs of pancreatic cancer years before diagnosis
According to the report, AI detects "invisible" signs of pancreatic cancer years before diagnosis. The report was published on 26 October.
doctorFor radiologists, this points to possible early detection and risk-based screening, though validation is still needed.drug development · globalPharma and AI firms sign $3.8 billion in deals in one month
AI–pharma agreements signed in October total $3.8 billion, and more than 70 alliances were formed in 2025. Among them, Pathos AI entered a development collaboration with Bristol Myers Squibb and…
pharmaceuticalsAI partnerships have become a standard business model; pipeline strategy must decide the balance between in-house and external AI capability.drug development · United StatesFormation Bio launches $600 million fund to buy clinical-stage drug candidates
Formation Bio has raised a $600 million fund to acquire clinical-stage drug candidates and finance their trials. The company is betting that AI-cheapened research will produce many candidates and…
pharmaceuticalsClinical trial capacity and the price of phase I–II candidates may rise in value, and exit options for smaller developers may improve.