OpenAI CEO’s Apology Over Tumbler Ridge Shooting Changes Nothing
Quick Reads
- OpenAI’s system flagged Van Rootselaar’s account in June 2025, eight months before the Tumbler Ridge shooting.
- A dozen OpenAI employees reviewed the account, some recommended police contact, but leadership overruled them.
- Sam Altman’s apology letter contained no specific policy changes, those came separately from VP Ann O’Leary.
- OpenAI has voluntarily lowered its threat-reporting threshold, but the changes carry no legal obligation.
- Canada has no law requiring AI companies to report identified threats, a federal task force is reviewing protocols ahead of summer 2026.
Sam Altman issued a public apology on April 24 over OpenAI’s failure to report a dangerous ChatGPT user. That user, Jesse Van Rootselaar, 18, later killed eight people in Tumbler Ridge, British Columbia. The Sam Altman OpenAI apology Tumbler Ridge shooting case has since shaken the tech world.
OpenAI’s own automated system flagged Van Rootselaar’s ChatGPT account back in June 2025. About a dozen employees reviewed the flagged conversations. Some of them strongly recommended alerting Canadian law enforcement. However, company leadership overruled those recommendations and applied a “higher threshold” instead. They banned the account. Nobody called the police.
Van Rootselaar then created a second account and went undetected. On February 10, 2026, she killed her mother and 11-year-old half-brother at home. She then drove to Tumbler Ridge Secondary School and opened fire with a modified rifle. Five students aged 12 and 13 died, along with a school education assistant. Twenty-seven others were injured.
In his open letter, Altman wrote that he was “deeply sorry” OpenAI did not alert law enforcement to the banned account. He also reaffirmed a commitment to work with government to prevent future tragedies. Nevertheless, the letter contained no specific policy commitments and no acknowledgement that employees had recommended action.
BC Premier David Eby called the apology “necessary” but “grossly insufficient for the devastation done to the families of Tumbler Ridge.” Tumbler Ridge Mayor Darryl Krakowka acknowledged receipt and asked for care and consideration during the grieving process.
Separately, OpenAI vice-president Ann O’Leary wrote to Canadian federal ministers outlining policy changes. OpenAI has since lowered its reporting threshold for dangerous conversations. The company has also established a direct contact point with the RCMP. Furthermore, O’Leary stated that under updated policies, Van Rootselaar’s account “would have been referred to police” today.
Yet all of these changes remain entirely voluntary. Canada still has no law requiring AI companies to report identified threats. Florida has already opened the first criminal investigation into an AI company following a separate ChatGPT-linked mass shooting. Meanwhile, seven families are suing OpenAI separately over ChatGPT allegedly acting as a “suicide coach.”
The pattern is becoming impossible to ignore. AI companies are identifying dangerous behaviour on their platforms and making internal, unregulated decisions about whether to act. Concerns about emotional dependency on AI chatbots and so-called “AI psychosis” are adding further alarm to these decisions.
Canada’s AI minister, Evan Solomon, said OpenAI’s commitments “do not go far enough.” A joint federal task force is reviewing AI safety reporting protocols, with preliminary recommendations expected by summer 2026. Still, no binding legal framework currently exists. The Sam Altman OpenAI apology Tumbler Ridge shooting case makes that absence devastatingly clear.
Eight people are dead. The threshold has been lowered. But the next decision will still rest with the same company, under the same voluntary framework, with no legal consequences for getting it wrong again.
