OpenAI’s chief executive has formally apologized to a small Canadian town where eight people were killed in a February mass shooting, saying the company should’ve gone to police sooner about the suspect’s activity on ChatGPT. In a Thursday letter to residents of Tumbler Ridge, British Columbia, Sam Altman wrote that he was “deeply sorry” that OpenAI didn’t alert authorities after suspending the user account of 18-year-old Jesse Van Rootselaar last June, reports the Wall Street Journal.
According to earlier reporting in the Journal, OpenAI’s automated systems had flagged Van Rootselaar’s messages for violent content, leading to an internal ban. Multiple staffers reportedly viewed the writings as a possible sign of danger and urged management to contact Canadian law enforcement, but the company declined to do so at the time. Once police publicly named Van Rootselaar after the shooting, OpenAI says it identified a second ChatGPT account tied to the suspect.
In his letter, first published by local outlet Tumbler RidgeLines, Altman offered condolences to the community and said he was renewing a pledge to the town’s mayor and British Columbia Premier David Eby to look for ways to prevent similar incidents. OpenAI says it has since strengthened its safety policies and that, under its updated rules for referring users to law enforcement, Van Rootselaar’s account would be reported if it were flagged today. Eby said this week that police are in the last phase of their investigation into the shooting.
“The apology is necessary, and yet grossly insufficient for the devastation done to the families of Tumbler Ridge,” Eby wrote on Facebook, per the AP. Tumbler Ridge Mayor Darryl Krakowka recently told reporters that residents are still grieving, with some only beginning to recover. The incident has also fed into Canada’s broader debate over artificial intelligence: Members of the governing Liberal Party recently backed a nonbinding resolution to bar AI chatbots for kids under the age of 16, as federal officials say they’re still weighing potential AI regulations. (Florida has opened a criminal probe into ChatGPT over a deadly shooting last year at Florida State University.)