OpenAI has pledged to report users whose conversations with ChatGPT “cross the line of imminent and probable danger” to the police. The company is taking these measures after a school shooting in Canada that killed eight people, according to The Washington Post.
The company sent a letter to the Canadian government outlining several immediate steps it had taken in response to the killings and noting that if these steps had been taken earlier, the police could have prevented the tragedy. Earlier, it was reported that OpenAI had decided not to report suspicious messages to law enforcement agencies.
This refers to the worst school shooting in Canadian history, which took place on February 10 in Trembler Ridge, British Columbia. Jesse Van Ruytseelaar killed eight people and then took her own life. Prior to this, OpenAI had blocked her ChatGPT account, but she was able to circumvent the restrictions by creating another account, which was discovered after the incident.
The first account was blocked in June 2025 for problematic use of the service, but at that time, OpenAI did not identify any “probable and imminent plans” that would have prompted the company to contact the police. After that, Rutselaar somehow managed to bypass the startup's restrictions and create another account, which was only discovered when the shooter's name was revealed — it was immediately reported to the police.
OpenAI's vice president of global policy, Anne O'Leary, said the company would strengthen its protocols for communicating with the police “when conversations cross the line of imminent and probable threat.” She also said that under the new expanded protocol for referring cases to law enforcement, Rutselaar's account, which was blocked in June, would have been immediately reported to the police if it had been discovered today.