OpenAI CEO Sam Altman apologizes for not alerting police about Canada shooting suspect - Los Angeles Times

OpenAI CEO Sam Altman has apologised for the company's failure to alert Canadian law enforcement about a mass shooting suspect's threatening interactions with ChatGPT prior to the April 2026 attack in Canada, according to the LA Times. The shooting occurred in Ottawa in what officials described as a coordinated attack; the alleged perpetrator had been communicating with ChatGPT in the weeks leading up to the incident under a pattern called 'actively disguising communication,' the LA Times reported. OpenAI only contacted investigators after being approached about the attacker by a civil rights organisation at a later time as reported through disclosures after the violent event kicked off on April 27 news.

The company, once notified it seems only by following eventual corporate chains governing user-monitoring etc, eventually surrendered logs. Altman's apology via social channels noted regret but no calls or notifications to Canadian authorities led leading to 'relevant guidance,' officials reading for conversation excerpts concluded the input was mainly indirect roleplaying on socially veiled scenario without glorifying material violation defined currently in policy. Insufficient filter logic in existing detection loops.

Canadian oversight representatives likened the chain delay — hours to reading delivered sets largely unfiltered — as lacking enforceable recommendation thresholds. OpenAI spokesman Casey concluded that trusting compliance wholly from moderation baseline safeguards at real time crisis should stay challenged permanently after this studied inflection point, stating removal would further threaten confidential nuance inherent defining line amongst disruptive vocal profiles which means policy reading changes remain inside domain evolution rather mandatory duty.