HeadlinesBriefing favicon HeadlinesBriefing.com

OpenAI CEO Apologizes For AI Safety Failure

Engadget •
×

OpenAI CEO Sam Altman has formally apologized to the Tumbler Ridge community for failing to report concerning ChatGPT conversations linked to the shooting suspect, Jesse Van Rootselaar. The apology comes two months after the deadly incident in British Columbia, with Altman acknowledging in a letter that the company banned the account in June for violating usage policies regarding potential real-world violence.

The chatbot exchanges raised red flags before the shooting, yet OpenAI never notified authorities. Altman explained he delayed making a public statement to "respect the community as you grieved" after speaking with both Tumbler Ridge Mayor Darryl Krakowa and British Columbia Premier David Eby. The tech giant now faces criticism for its delayed response to potential threats that might have prevented the tragedy.

Premier Eby acknowledged the apology was "necessary" but called it "grossly insufficient" for the devastation suffered by families. Altman reaffirmed OpenAI's commitment to work with governments to prevent similar tragedies, building on previous policy announcements from vice president of global policy Ann O'Leary about reporting "imminent and credible" threats found in user conversations. The company faces increased scrutiny over its responsibility to monitor for dangerous content.