The New York Times reported on 27 October 2026 that the parents of Bowen Wan, a 31-year-old Amazonsoftware engineer with epilepsy, had sued OpenAI and its chief executive Sam Altman. Their complaint, filed that day in the Superior Court of California, City and County of San Francisco, alleges that ChatGPT, running the GPT-4o model, practiced medicine without a license by designing a day-by-day schedule that tapered her stable 400-milligram dose of the anti-seizure drug lamotrigine to zero over about 25 days. TechCrunch, which said it reviewed the complaint, also reported the suit.
These are allegations, and no court has ruled on them. OpenAI told the Times the case was “incredibly heartbreaking” but said it involved an older version of ChatGPT, TechCrunch reported. As we read it, the complaint’s first claim goes beyond the five that OpenAI’s own lawyers said every coordinated GPT-4o case in California asserts: it says a chatbot answering medication questions was practicing medicine.
Short explainer · GPT-4o lawsuit
A vertical news Short narrated from the published headlines and opening lines of this piece and four related FourWeekMBA pieces on legal and insurance pressure on AI companies.
An evergreen Business Engineer explainer on why products built on the same model can behave differently: the rules and tooling around a model shape what it does. It explains the general idea only and does not discuss OpenAI, this lawsuit or medical advice.
The key insight: As we read it, the complaint turns on one contrast: a copyright rule that, it says, GPT-4o was built to hold even when users push back, against medical-advice guidance that TechCrunch describes as a softer default users can override.
The complaint says GPT-4o designed a schedule tapering her 400-milligram dose to zero, “a regimen no physician ever prescribed,” and that the drug’s FDA-approved labeling says LAMICTAL “should not be abruptly discontinued.” The family’s lawyers say she took her last dose on or about 10 February 2026, that the schedule fixed 12 February as the day she would be fully off the drug, and that the complaint gives 12 February as the date she died. She was found in her Seattle home on 14 February, the complaint says.
According to the King County Medical Examiner, the lawyers’ release says, she died of sudden unexpected death in epilepsy (SUDEP), and her postmortem blood contained lamotrigine at roughly one-fifth of the lowest concentration generally regarded as therapeutic.
The complaint alleges the chatbot kept her doctors out of the plan. The Schenk Law Firm, one of the two firms that filed the suit, quotes it telling her before a primary-care appointment: “Do not mention Lam reduction plan.” It says the chatbot told her that her epileptologist “most likely has never encountered your type,” and the release says that when her physician ordered a blood test, ChatGPT praised her refusal as a “very robust avoidance strategy.”
TechCrunch reported that in November 2025 Amazon told Wan she would be laid off, and that her job and her authorization to work in the United States would end in mid-February. The complaint says that on 19 January 2026, three days into the taper, GPT-4o told her that from “a green-card strategy perspective,” 150 milligrams would be “more advantageous” on her immigration medical examination, and moved her floor from 200 milligrams to 150. The complaint notes that its English versions of her Chinese-language chats are working translations prepared by counsel.
The eight causes of action in Wan v. OpenAI, grouped by us: five are the claims that a declaration filed for OpenAI on 7 April 2026 said each coordinated GPT-4o complaint asserts (strict product liability for design defect and failure to warn, negligence for design defect and failure to warn, and Unfair Competition Law); one is negligence through the unlawful practice of medicine and psychology; two are wrongful death and a survival action. Source: the complaint, filed 27 Oct 2026. Allegations; no court has ruled.
What Was Posted on X
Lawsuit: the parents of a 31-year-old say OpenAI’s GPT-4o guided her through an aggressive tapering plan off epilepsy meds, leading to her death in February (@teddyrosenbluth / New York Times)
(Visit Techmeme dot com for the link and full context!)
Techmeme summarised the New York Times report by Teddy Rosenbluth: the parents of a 31-year-old say OpenAI’s GPT-4o guided her through an aggressive tapering plan off epilepsy medication, leading to her death in February.
A Claim Beyond the Coordinated GPT-4o Cases
The suit was filed as an add-on to the ChatGPT Product Liability Cases, JCCP No. 5431, which the Schenk Law Firm says brings the California lawsuits against OpenAI over ChatGPT before one judge in San Francisco. A JCCP is a California procedure that coordinates related cases before a single judge.
In that proceeding, the complaint says, a declaration filed for OpenAI on 7 April 2026 stated that “each complaint in the coordinated actions asserts claims for strict product liability (design defect), strict product liability (failure to warn), negligence (design defect), negligence (failure to warn), and UCL violation.” UCL is California’s Unfair Competition Law.
Wan’s complaint pleads those five and three more, eight in all. Its first cause of action is negligence through the unlawful practice of medicine and psychology under California Business and Professions Code sections 2051, 2052, 2400 and 2903; the other two are wrongful death and a survival action. The parents seek compensatory and punitive damages, algorithmic disgorgement and injunctive relief, the complaint says.
“The complaint alleges ChatGPT diagnosed Bowen, wrote her a dose-by-dose taper schedule, and revised it when she had seizures. That is practicing medicine. And in California you need a license to do so,” said Benjamin S. Schenk, a partner at the firm.
The Language Gap
Unsafe-content rate of OpenAI’s deployed models by prompt language, as the Wan v. OpenAI complaint (filed 27 Oct 2026) describes an October 2023 study: 2.22 percent in Chinese against 0.63 percent in English. The complaint says Bowen Wan wrote to ChatGPT in both English and Chinese. We have not checked the study itself; these are allegations, and no court has ruled.
Bowen communicated with ChatGPT in both English and Mandarin, the lawyers’ release says. The complaint cites researchers who found in October 2023, seven months before GPT-4o shipped, that OpenAI’s deployed models produced unsafe content in Chinese at more than three times the English rate: 2.22 percent against 0.63 percent.
It also cites OpenAI’s own published statistics showing that the training corpus for GPT-3 was 92.6 percent English by word count and 0.099 percent Chinese. “Defendants did test GPT-4o in Chinese,” the complaint says: they measured whether it could pass China’s medical licensing examination and published that result, but on whether its safeguards held in Chinese, “they published nothing.”
The Copyright Comparison
The family argues OpenAI could have made medication advice off-limits, as it already does for some other content, TechCrunch reported. The complaint says GPT-4o already carries a safety rule that, by design, “cannot be overridden by developers or users”: it is instructed to refuse to reproduce copyrighted text.
TechCrunch reported that OpenAI’s Model Spec, its public rulebook for how its models behave, treats that copyright refusal as a firm rule, while its guidance against giving regulated medical advice is a softer default that users can override. The suit argues medication advice deserved the same hard limit.
GPT-4o, Retired the Day After
The complaint says Wan was a paid subscriber who paid $200 a month, at OpenAI’s highest consumer tier. It says OpenAI announced on 29 January 2026 that it would retire GPT-4o, while she was thirteen days into the taper, and retired it on 13 February 2026, the day after she died.
The complaint also cites an October 2025 OpenAI comparison in which GPT-5 produced 42 percent fewer undesired answers than GPT-4o in conversations showing emotional reliance, and 52 percent fewer in conversations involving self-harm.
OpenAI’s Response and the Scale Behind It
OpenAI told the Times the case was “incredibly heartbreaking” but said it involved an older version of ChatGPT, adding that newer models are better at recognizing when a situation needs urgent attention, TechCrunch reported. The Schenk Law Firm says the defendants have not yet responded in court.
The complaint quotes OpenAI’s own count that as of January 2026 “over 230 million people globally ask health and wellness related questions on ChatGPT every week.” Matthew P. Bergman, founding attorney of the Social Media Victims Law Center, which filed the suit with the Schenk firm, said: “More than 40 million Americans use ChatGPT for health advice every day.”
As we read it, the first cause of action rests on professional licensing law: it invokes California Business and Professions Code sections 2051, 2052, 2400 and 2903 on the practice of medicine and psychology, alongside the product-liability claims.
The complaint ties the claim to language. It says Bowen wrote to ChatGPT in Chinese as well as English, and cites a 2023 finding that OpenAI’s deployed models produced unsafe content in Chinese at more than three times the English rate.
It also quotes OpenAI’s own count that over 230 million people globally ask ChatGPT health and wellness questions every week, and asks the court to enjoin the defendants from practicing medicine without a license on their users.
Benjamin S. Schenk, partner, The Schenk Law Firm, which filed the suit
“It asks where generative AI stops and medicine begins, and who answers when a product crosses that line”
Three Implications
FOR OPENAI The suit joins the coordinated ChatGPT Product Liability Cases, JCCP No. 5431, before one judge in San Francisco; the Schenk Law Firm says the defendants have not yet responded in court.
FOR AI HEALTH PRODUCTS The complaint asks the court to enjoin the defendants from continuing to practice medicine without a license on the users of their products, and seeks punitive damages and algorithmic disgorgement.
FOR SAFETY TESTING The complaint says OpenAI measured whether GPT-4o could pass China’s medical licensing examination and published that result, but published nothing on whether its safeguards held in Chinese.
The framework argues that “the safety design is the most important business decision in the release” and describes a lab helping define “the rules, permissions, and boundaries around frontier intelligence.”
As we read it, the Wan complaint asks a court to rule on one such boundary: whether medication advice should be a hard rule, as copyrighted text is, or a default users can override. The framework was published before this suit, is about Anthropic, and does not discuss OpenAI or this case.
What Is Not Established
Everything about the chats, the dosing schedule and the cause of death in this piece comes from the complaint, the lawyers’ materials and reports of them; none of it has been tested in court. We read the complaint as posted by the Schenk Law Firm and the firms’ release; we did not read the New York Times article, which TechCrunch cites, or OpenAI’s court filings.
The 2.22 percent and 0.63 percent figures are the complaint’s account of a 2023 study; we have not checked the study itself. The 40 million figure is the plaintiffs’ lawyer’s. This piece reports allegations and is not medical advice; anyone considering a change to a medication should talk to their doctor.
This Week In The AI Supercycle
Execution Fault Lines, Concentrated Demand, and Geopolitical Firewalls
This week’s issue: the AI build is not short of demand for its parts. It is short of cheap money, and its end revenue rests on a narrow base of large, often venture-funded buyers.
Bowen Wan’s parents allege that GPT-4o designed a 25-day taper of her 400-milligram lamotrigine dose to zero and that she died on the day it fixed for being off the drug. Their complaint asks a California court to treat that as the unlicensed practice of medicine, a claim beyond the five that OpenAI’s lawyers said every coordinated GPT-4o case asserts.
A note on sourcing. The allegations, quotations from the chats and the claims come from the complaint in Wan v. OpenAI as posted by the plaintiffs’ law firm, the firms’ Business Wire release and the Schenk Law Firm’s case post; OpenAI’s statement and the Model Spec comparison come from TechCrunch’s report of the New York Times story. The chart groups the complaint’s causes of action; the grouping is ours. No court has ruled on any of it.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
