OpenAI Apologizes for AI-Generated Police Alert in BC Shooting

OpenAI CEO Sam Altman apologized on Friday, June 21, to the community of Tumbler Ridge, British Columbia, after the company did not notify law enforcement about a ChatGPT account linked to a school shooter ahead of a February 2025 mass killing. The apology was delivered in a letter that was shared publicly by British Columbia Premier David Eby.

Officials said eight people died in the February 2025 incident. The deaths included six at Tumbler Ridge Secondary School and two at a nearby residence, before the shooter—identified as 18-year-old Jesse Van Rootselaar—died by suicide.

Altman’s letter said Van Rootselaar’s ChatGPT account had been banned in June 2024, about eight months before the attack. OpenAI had previously stated that automated systems and human reviewers flagged the account for possible misuse connected to violent activity, and that the account was removed for violating usage policies.

Despite the ban, the company said it did not refer the matter to police at the time. OpenAI’s position, as described in the letter, was that the account did not reach the company’s threshold for contacting law enforcement because it was not assessed as an imminent and credible risk of serious physical harm.

The apology in British Columbia came as OpenAI also faced scrutiny in the United States. Earlier in the week, Florida Attorney General James Uthmeier opened a criminal investigation into OpenAI tied to allegations that ChatGPT provided “significant advice” to a student accused in an April 2025 campus shooting at Florida State University that left two people dead.

OpenAI said that, after learning of the Florida incident, it proactively provided information about the suspect’s ChatGPT account to law enforcement. The company’s statements describe two different responses across the two cases: a decision not to escalate to police in 2024 in British Columbia, and a decision to share account information after the April 2025 Florida shooting became known.

For governments, schools, and technology firms, the developments highlight how internal safety reviews, account enforcement, and referral standards can become central issues after violent incidents. The cases also show that questions about when a platform should alert authorities can quickly move from internal policy to public accountability and legal processes.

Key uncertainties remain based on the information provided. OpenAI has not detailed what specific content triggered the June 2024 flagging, nor what factors led reviewers to conclude the account did not meet the referral threshold at that time. The Florida investigation is also at an early stage, with the allegations described as part of an inquiry initiated by the state attorney general.