OpenAI employees raised alarms regarding violent chatbot prompts submitted by mass shooter Jesse Van Rootselaar months before he killed family members and school-aged children in Tumbler Ridge, British Columbia. According to a report by the Wall Street Journal, an automated review system originally flagged the interactions, which featured violent scenarios involving gun violence over several days. Around a dozen employees at the artificial intelligence organisation were aware of the ChatGPT activity, and some expressed a desire to contact law enforcement. Ultimately, OpenAI did not notify the police, adhering to its policy of only contacting authorities when there is an imminent threat of real-world harm or violence.

Background: Generative AI in Law Enforcement

← Back to the news feed