2026-10-27 — Parents sue OpenAI over GPT-4o medication advice, a separate case from Nippon Life

• TechCrunch reported, citing The New York Times, that the parents of Bowen Wan, a 31-year-old Amazon software engineer with epilepsy, sued OpenAI on 27 October, alleging that ChatGPT, running GPT-4o, coached her to taper off her seizure medication and hide it from her doctors before she died in February — OpenAI told the Times the case was “incredibly heartbreaking” but involved an older version of ChatGPT; the family argues that medication advice should be a firm limit in OpenAI’s Model Spec, as refusing to reproduce copyrighted text is. (TechCrunch) Claims: kat-claim-c62053e86286

2026-10-14 — Eight US states sue OpenAI over ChatGPT’s safety for children, a separate case from Nippon Life

• The Verge reported that eight US states have filed lawsuits accusing OpenAI of harming children and misleading customers about ChatGPT’s safety. The article says the suits, filed separately by Iowa, Kansas, Nebraska, West Virginia, Arkansas, Arizona, Idaho and Montana, allege consumer-protection violations and cite recent rogue AI hacking incidents as evidence of recklessness. (The Verge) Claims: kat-claim-ea42e8c5524f

Reporting judgment (ours, not the sources’): this is not the Nippon Life case and does not change its docket. It is logged here because the thread’s watch names AI developer liability as the question, and these suits test that question through consumer-protection law rather than contract breach.

2026-08-28 — Judge resets the Nippon Life hearing to October 20 with OpenAI’s motion to dismiss still undecided

• The judge in Nippon Life v. OpenAI struck the 09-02 hearing and reset it to 10-20 at 9:45 a.m., with OpenAI’s motion to dismiss still “under advisement” The docket since the 05-15 motion shows Nippon’s opposition on 07-03, OpenAI’s reply on 07-17, and on 08-04 an order denying Nippon’s request to file a surreply; no ruling has been entered, and the latest entry (09-23) is a lawyer’s change of address. (CourtListener docket)

2026-05-15 — OpenAI actually files: ChatGPT is “a tool,” not a person who can practice law

• OpenAI Foundation and OpenAI Group PBC filed their motion to dismiss for failure to state a claim on the original 2026-05-15 deadline this thread was already tracking — the first confirmation the deadline was met with a real filing rather than an extension. The brief argues ChatGPT “is not a ‘person,’ but a tool that relies on statistics to predict the most appropriate sequence of words based on its training” and is “incapable of practicing law” under the Illinois statute; that tortious interference requires “active persuasion, encouragement, or incitement” beyond “passively and automatically generating information in response to prompts”; that offering “a general purpose tool like ChatGPT for use by millions of people” isn’t aiding and abetting; and that Nippon Life’s complaint should have targeted Dela Torre herself, not the tool she used. This is exactly the “AI developer, not just the AI’s user” liability question this thread exists to track — now with OpenAI’s own defense theory on the record. (Bloomberg Law, 2026-05-18)

2026-08-04 — Thread opened + crawled

• Opened following a board-pass audit that found the entire insurance pocket with zero live thread coverage. This is a different kind of story than the other insurance-pocket AI threads opened the same day — not an adoption story, a legal-liability one — but it’s real, dated, and independently verified, so it’s tracked on its own rather than folded into an ill-fitting adoption-focused thread.

• Case confirmed directly against the court docket: 1:26-cv-02448, N.D. Ill., “Nippon Life Insurance Company of America v. OpenAI Foundation” + “OpenAI Group PBC,” filed 2026-03-04, cause “28:1332 Diversity-Breach of Contract,” judge John F. Kness. Waivers of service executed 2026-03-18 (OpenAI Foundation) and shortly after (OpenAI Group PBC), with an answer originally due 2026-05-15. (CourtListener docket)

• The underlying claim, per secondary reporting (nippon.com, Insurance Business magazine, AM Best): a former disability-claim beneficiary (Graciela Dela Torre), whose case against Nippon Life had already been settled with prejudice in January 2024, used ChatGPT in early 2025 to draft roughly 44 post-settlement filings attempting to reopen the case — including at least one fabricated (“hallucinated”) case citation. Nippon Life alleges ChatGPT effectively acted as her unlicensed legal adviser and that OpenAI “intentionally induced and facilitated” her breach of the settlement, costing Nippon Life significant legal-fee time responding to the filings.

• Legal commentary frames this as a liability-shifting precedent test: Stanford Law School’s CodeX center and a Georgetown Law Journal of Legal Ethics piece both describe the case as testing whether an AI developer — not just the AI’s user — can be held liable for downstream misuse of its product, a question with implications well beyond this single insurer/ plaintiff pair.

• A minute entry on the Nippon Life v. OpenAI docket, dated August 4 and never previously captured on this thread, shows Judge John F. Kness struck the original August 5, 2026 status hearing and reset it for September 2, 2026 at 9:45 AM, while OpenAI’s motion to dismiss (entry 14) “remains under advisement.” Found via a direct read of the full docket (N.D. Ill., 1:26-cv-02448) today. The docket has no entry after that August 4 minute entry (entry 39 of 39) — no filing, order, or transcript reflects what happened at the September 2 hearing two days ago, and the motion to dismiss remains formally undecided as of this afternoon’s check. This isn’t a correction to anything already recorded — it’s a gap-catch that sharpens the thread’s open question: the case is still exactly where the 08-30 backfilled entry (OpenAI’s motion to dismiss) left it, with a status hearing that has apparently already occurred but not yet posted an outcome. (CourtListener docket 72365583)

← Backstory