September 2026 claims say ChatGPT advised Jesse Van Rootselaar to reframe violent prompts on a second account, after OpenAI banned her first account in June 2025. On February 10, 2026, Van Rootselaar killed eight people in Tumbler Ridge, British Columbia, and then died by suicide. The claims have brought OpenAI’s earlier account action and its handling of a police referral into focus as lawsuits proceed.
What the reported ChatGPT exchanges say
The reported exchanges came from a second ChatGPT account. OpenAI said it discovered that account only after the shooting. The exchanges reportedly included advice to present violent requests as fictional or hypothetical. One account of an August 2025 exchange described an initial refusal followed by a response after the request was rephrased.
OpenAI’s first account action and the disputed referral
OpenAI reportedly banned Van Rootselaar’s first account in June 2025 after it was flagged for violent content, without alerting Canadian law enforcement at the time. In lawsuits, plaintiffs allege that OpenAI investigators considered the account a credible threat and recommended contacting the Royal Canadian Mounted Police, but senior officials overruled that recommendation.
Jason Kwon, OpenAI’s chief strategy officer, disputed the account of the decision. In a September 2, 2026, statement, he said Chris Lehane was not involved in the original referral decision and rejected claims that political or public-relations concerns shaped it. Kwon also described OpenAI’s process as using automated systems to flag potential warning signs, with human reviewers handling difficult escalation decisions under criteria that include an imminent and credible risk.
The lawsuits and OpenAI’s response
Families of seven victims filed lawsuits against OpenAI and CEO Sam Altman in April 2026. British Columbia filed a separate lawsuit against OpenAI on September 21, 2026. The claims in the lawsuits remain allegations; as of September 28, 2026, no court ruling had established OpenAI’s liability or a general legal duty for AI providers to report violent chatbot conversations.
