British Columbia has sued OpenAI and chief executive Sam Altman, alleging that the company failed to alert police after detecting ChatGPT conversations connected with the person who later committed the February 2026 Tumbler Ridge school shooting.
The complaint, filed in a California federal court on 21 September, alleges that OpenAI’s safety team flagged the account’s discussions of gun violence and recommended contacting police, but management instead deactivated the account. It further alleges that the user created another account and continued using ChatGPT while planning the attack.
The province seeks damages for recovery costs—including replacement of the school and mental-health services—and a court order requiring changes to OpenAI’s handling of conversations indicating potential violence. The allegations have not been determined. OpenAI says the original activity did not satisfy its reporting threshold and that it remains committed to working with governments and law enforcement.
The case is significant because it moves AI liability beyond harmful answers generated by a model. It asks whether a provider that actively monitors conversations may acquire a duty to escalate credible warnings—and whether inadequate thresholds, account-linking or internal decision processes can ground negligence or product-liability claims.
For Australian lawyers and regulators, the difficult questions will include foreseeability, causation, privacy, confidentiality and when safety monitoring becomes a legally enforceable responsibility.
Sources: Reuters on the lawsuit; OpenAI’s earlier safety response.
Leave a Reply