Sam Altman, CEO of OpenAI, issued a public apology to residents of Tumbler Ridge, Canada, a town marked in February by a fatal attack by a young woman who had previously posted alarming messages in chats with ChatGPT.
The letter, published on Friday by the local publication Tumbler RidgeLines, emphasizes that “no one should have to go through a tragedy like this”, according to News.ro
Altman declares “deeply grieved” that the company did not alert the authorities about the discussions the author of the attack had on the platform.
The account of Jesse Van Rootselaar (18) had been suspended eight months before the massacre, after his activity raised suspicions of violent intentions. But OpenAI stressed that at the time there were no indications of imminent danger, which is why the police were not notified, according to information cited by AFP.
“No explanation can repair the losses suffered”
In February this year, a transgender woman killed her mother and half-brother, then opened fire in a former school, seriously injuring five children and a teacher, before killing herself.
In the letter to the community of Tumbler Ridge, Altman states that although “no explanation can repair the losses suffered”, considers it necessary to take moral responsibility for the fact that the discussions were not reported to the authorities.
OpenAI is facing intense criticism in Canada, and company officials have been summoned to Ottawa to detail safety protocols and emergency response procedures.
The family of one of the injured victims filed a complaint against the company in March, accusing the lack of adequate preventive measures.
We recall that OpenAI admitted in February that the 18-year-old transgender teenager who killed her family and several people in a school in Canada had talked to ChatGPT in the year before the attack and had shown a predisposition to violent behavior. Although the company considered notifying the authorities, it ultimately decided not to notify the police because there were no indications of an imminent attack.
According to an official statement, cited by The Guardian, OpenAI analyzed the activity of Jesse Van Rootselaar’s account in June 2025, a few months before she committed one of the worst armed attacks in recent Canadian history.
The account has been identified by internal abuse detection systems and has been flagged for “supporting or facilitating violent activities”.
