In response to a recent school shooting, OpenAI announced on Thursday that it will establish a direct line of communication with Canadian law enforcement and better identification of serial offenders of its “violent activities” policy in order to strengthen safety standards. The ChatGPT creator outlined the process in a letter to Evan Solomon, Canada’s minister in charge of artificial intelligence.
Ann O’Leary, OpenAI’s vice president of global policy, wrote the letter after Canadian ministers encouraged the ChatGPT creator to improve its safety standards as soon as possible, warning Ottawa that if it did not, the firm would face legislative action. “We remain committed to cooperating with law enforcement authorities in the investigation into the Tumbler Ridge tragedy, and we are committed to an ongoing partnership with the federal and provincial governments,” O’Leary said, referring to the town in British Columbia where the shooting took place.
Ottawa summoned OpenAI’s safety team for meetings this week after the business stated that it had not alerted police about the account of the alleged shooter, Jesse Van Rootselaar, which it had suspended. Van Rootselaar, 18, is suspected of murdering eight people on February 10 before killing herself in Tumbler Ridge. According to OpenAI, her ChatGPT account was banned last year due to rules infractions.
The company stated that the account was detected by technologies that detect “misuses of our models in furtherance of violent activities,” but did not disclose any information. OpenAI stated that the issues did not fulfil its internal standards for reporting to law authorities.
O’Leary stated on Thursday that under the company’s “enhanced law enforcement referral protocol,” if the initial account ban in June had been discovered now, it would have been reported to authorities. The company learned Van Rootselaar utilised a second account, which was disclosed with law authorities. “We commit to strengthening our detection systems to better prevent attempts to evade our safeguards and prioritise identifying the highest-risk offenders,” O’Leary added.
The corporation also agreed to periodically evaluate the levels set by its automated algorithms to detect potential violent behaviour by users. Crime experts have stated that, while increased surveillance of AI platforms and social media is required, police or other authorities may have missed additional opportunities to prevent one of Canada’s deadliest mass killings. Police claimed Van Rootselaar had a history of mental health issues, and they had taken and later returned guns from her house. Minister Solomon’s office did not immediately reply to a request for comment.
