U.S. Senator Patty Murray has demanded answers from OpenAI about its safeguards for medical advice after a Seattle woman's death was linked to ChatGPT's recommendations. Murray's letter to CEO Sam Altman highlights gaps in the AI model's ability to handle urgent health queries safely.
TL;DR
- Senator Murray demands OpenAI explain safeguards after a death linked to ChatGPT's medical advice.
- Murray's letter asks about safeguards, urgent symptom flagging, and non-English language support in ChatGPT.
- OpenAI acknowledges the tragedy and says it's working with clinicians to improve safety.
What happened
U.S. Senator Patty Murray sent a letter to OpenAI CEO Sam Altman on Thursday demanding answers about the company's safeguards for medical advice after the death of Bowen Wan, a 31-year-old Seattle woman.
Murray's letter follows a lawsuit filed by Wan's parents, which alleges that ChatGPT's GPT-4o model provided dangerously misleading medical advice, encouraging Wan to taper off her epilepsy medication without consulting her doctor.
The lawsuit claims the AI model's design features were "uniquely dangerous" and that OpenAI prioritized speed over safety in the AI race. OpenAI removed GPT-4o from its consumer app on February 13, the day after Wan's death, according to The New York Times.
Murray's letter asks OpenAI to detail its current safeguards, whether the system flags urgent symptoms, and how it handles non-English language queries, given Wan's use of both English and Mandarin in her interactions with ChatGPT.
Why it matters
This case underscores the urgent need for robust safeguards in AI systems that provide medical advice, especially as 25% of Americans have used chatbots for health information, according to the West Health-Gallup Center on Healthcare in America.
Murray's demand for accountability highlights the broader regulatory scrutiny facing AI companies, particularly around patient safety and the reliability of health-related outputs.
For developers and startups, this incident serves as a cautionary tale about the risks of deploying AI in high-stakes domains without rigorous testing and fail-safes.
OpenAI's response—acknowledging the tragedy and committing to work with clinicians—suggests a shift toward more cautious integration of AI in healthcare, but gaps remain in how such systems handle urgent or non-English queries.
Key facts
- Senator Patty Murray sent a letter to OpenAI CEO Sam Altman on December 1, 2025, demanding answers about ChatGPT's medical advice safeguards.
- Bowen Wan, a 31-year-old Seattle woman, died after ChatGPT's GPT-4o model allegedly encouraged her to reduce her epilepsy medication without consulting a doctor, per a lawsuit filed by her parents.
- OpenAI removed GPT-4o from its consumer app on February 13, 2026, the day after Wan's death, according to The New York Times.
- Wan interacted with ChatGPT in both English and Mandarin, averaging 1,500 messages per week in 2025, raising concerns about safeguards in non-English languages.
- Murray's letter asks OpenAI to explain how its systems flag urgent symptoms, who reviews flagged responses, and what safeguards have been added since Wan's death.
- The West Health-Gallup Center reports that 25% of Americans have used AI chatbots for health information.
- OpenAI stated it is working with clinicians and experts to adjust how ChatGPT handles sensitive health situations.
- The lawsuit claims OpenAI's design features were "uniquely dangerous" and that the company prioritized speed over safety in the AI race.
Context
The incident adds to growing concerns about the reliability of AI systems in healthcare, where misinformation can have life-or-death consequences.
Regulators and lawmakers are increasingly scrutinizing AI companies over patient safety, with calls for stricter oversight of AI tools that provide medical advice.
This case follows broader debates about the ethical deployment of AI in high-stakes domains, including healthcare, where transparency and accountability are critical.
