Las Vegas police have revealed that Matthew Livelsberger, a Green Beret veteran who took his life after detonating a Tesla Cybertruck near Trump Towers, used ChatGPT and other generative AI tools to aid his attack planning. The incident marks the first known case on U.S. soil where a generative AI platform was allegedly used in preparing for a violent act.
According to Las Vegas Metropolitan Police Department sheriff Kevin McMahill, a review of Livelsberger’s ChatGPT queries showed searches on explosive targets, bullet speeds, and fireworks legality. “It’s a concerning moment,” McMahill said during a press conference, emphasizing the troubling intersection of AI tools and violent behavior.
OpenAI, the company behind ChatGPT, expressed regret over the incident and confirmed it is cooperating with the investigation. “Our models are designed to refuse harmful instructions and minimize harmful content,” an OpenAI spokesperson said, adding that the information Livelsberger accessed was publicly available, with warnings against harmful activities provided by the platform.
Livelsberger, an Afghanistan veteran, detonated the Cybertruck using improvised explosives and fireworks, injuring seven bystanders but causing no fatalities other than his own. Authorities believe Livelsberger was suffering from PTSD. His note revealed disillusionment with America’s future and bizarre comments about drones with “gravitic propulsion systems.” He also expressed support for Tesla CEO Elon Musk and President-elect Donald Trump.
The incident raises concerns about the misuse of AI tools like ChatGPT for harmful purposes. While AI companies have implemented guardrails to prevent misuse, incidents like this highlight the challenges in enforcing those protections and underscore the need for vigilance as AI technologies become more widespread.
