What Happened
OpenAI's internal safety team identified a ChatGPT account linked to a future school shooter as posing a credible gun violence threat more than eight months before the attack in Tumbler Ridge, Canada. The company was supposed to alert police, who already had a file on the shooter and had previously removed guns from their home. Instead, OpenAI simply deactivated the account and told the user how to sign up again with a different email address.
Why You Should Care
If AI companies won't report credible threats because it might hurt their stock price, your kids' school safety depends on corporate earnings calls.
📚 The Basics
AI safety teams are supposed to monitor for users planning real-world violence and report credible threats to law enforcement. Companies like OpenAI have internal protocols for this, but they're not legally required to report in most jurisdictions. When a company goes public (IPO), any scandal or liability can tank the stock price and cost billions. Canada has had relatively few mass school shootings compared to the US, making this incident particularly shocking for the small mining town of Tumbler Ridge.
🧠 Look Smart At Dinner
Say This
The really damning part is they didn't just ignore the threat — they literally gave the shooter instructions on how to get back on the platform.
Context
OpenAI told the banned user exactly how to circumvent their safety measures by signing up with a different email address.
Avoid Saying
Don't say 'AI can't predict human behavior' — their own safety team specifically flagged this user as a credible threat months in advance.
The Approved Opinion™
“Technology companies have a responsibility to balance user privacy with public safety, and this case shows we need clearer guidelines for when to involve law enforcement.”
