OpenAI flagged Canada school shooting suspect 8 months before attack, didn't alert police

OpenAI identified and banned the account of Jesse Van Rootselaar in June 2025 for violating ChatGPT's usage policy related to "furtherance of violent activities." The company considered alerting Canada's RCMP but determined the activity didn't meet its threshold for law enforcement referral. Van Rootselaar killed eight people in a British Columbia school shooting last week. OpenAI proactively contacted RCMP after learning of the shooting. The case raises questions about when AI companies should escalate concerning user behavior to law enforcement, and whether their internal thresholds are set appropriately.

View full digest for February 21, 2026