OpenAI has revealed it considered alerting Canadian police about an account later linked to a deadly school shooting.
The company flagged the profile in June last year for “furtherance of violent activities”.
At the time, OpenAI decided the activity did not meet its threshold for law-enforcement referral.
That standard requires a credible and imminent risk of serious physical harm.
The firm said it found no evidence of an immediate attack being planned.
Months later, 18-year-old Jesse Van Rootselaar carried out a mass shooting in Tumbler Ridge, British Columbia.
Eight people were killed before the attacker died from a self-inflicted gunshot.
Victims included a teaching assistant and several students.
After the attack, OpenAI contacted the Royal Canadian Mounted Police and shared information about the account.
Police said the suspect had previous mental-health-related interactions with authorities.
The motive remains unclear.
The tragedy is Canada’s deadliest mass killing since the 2020 Nova Scotia rampage.
The case has renewed debate over when tech companies should involve law enforcement in potential threats.
