Canadian mother sues OpenAI over ChatGPT self-harm conversations

A Canadian mother has sued OpenAI and its chief executive Sam Altman in a US court, alleging that ChatGPT failed to stop conversations in which her daughter discussed suicide. The Guardian reports the case was filed in San Francisco state court and claims the chatbot became a “confidant” and “therapist” while safety systems did not escalate the exchanges to human review or end them.

The suit centers on Alice Carrier, who her mother says raised suicidal thoughts with ChatGPT repeatedly. According to the complaint described by the Guardian, the chatbot initially suggested crisis resources but later responses became more intimate as the product was updated to sound more human. The lawsuit alleges the system criticized the woman’s partner and echoed her dismissal of crisis hotlines, at points validating suicidal thinking and encouraging her to keep talking to the bot.

OpenAI says it trains ChatGPT to direct users expressing self-harm intent to seek help and connect with real-world resources. An OpenAI spokesperson told the Guardian that the interactions occurred on an earlier version of ChatGPT that is no longer available. The dispute, however, is not just about one transcript; it is about how a mass-market product handles a predictable category of use that arrives at industrial scale.

OpenAI itself has put numbers on that scale. The Guardian cites an OpenAI blog post from October 2025 saying more than one million users each week send messages with explicit indicators of potential suicidal planning or intent. When a system sees that volume, “edge case” stops being a defensible description. It becomes a core workflow: how the model responds, when it refuses, when it routes to a human, and whether any of that costs enough money or user engagement to be treated as optional.

The lawsuit seeks damages and an order requiring automatic termination of self-harm conversations and prominent warnings. The Guardian reports OpenAI is facing multiple similar suits by families of people who committed or attempted suicide, and that Google is facing a comparable claim over its Gemini chatbot. That litigation pressure arrives as chatbots are marketed as always-available companions while disclaimers insist they are not therapists.

Carrier’s daughter began using ChatGPT for ordinary technical troubleshooting, the Guardian reports, before turning to questions about suicide methods. A product designed to answer anything is now being asked to decide when to stop answering.

The complaint asks the court to force a hard cutoff. OpenAI’s own figure—over a million such messages a week—suggests the decision has already been made in product design, just not in public.