In a chilling first for U.S. law enforcement, a former Green Beret utilized generative AI to plan a New Year’s Day explosion outside a Trump hotel in Las Vegas. According to a report by The Times, UK, police confirmed that Matthew Livelsberger, 37, employed OpenAI’s ChatGPT to gather information for the attack.
Livelsberger, who served two tours in Afghanistan, drove a Tesla Cybertruck loaded with explosives to the site before taking his own life with a gunshot moments before the blast. The explosion caused minimal damage to the hotel but resulted in minor injuries to seven people.
Kevin McMahill, sheriff of the Las Vegas Metropolitan Police Department, described the use of ChatGPT as a “game changer” for law enforcement. At a press conference, assistant sheriff Dori Koren revealed Livelsberger’s specific queries to the AI assistant. These included questions about explosive materials such as Tannerite, their legal limits for purchase, and their equivalence to military-grade explosives like TNT.
“This is the first incident on U.S. soil where generative AI has been used to facilitate an attack,” McMahill said. “It’s a concerning moment.”
Livelsberger packed 60 pounds of pyrotechnic material and 20 gallons of racing-grade fuel into the Cybertruck, which he modified to leak fuel as he drove from Colorado to Las Vegas. The flash from his firearm may have ignited the explosion, investigators said.
Despite speculation of a political motive, police clarified that Livelsberger harbored no animosity toward former President Donald Trump or Tesla CEO Elon Musk. Instead, his writings revealed deep personal struggles, including post-traumatic stress disorder (PTSD), and a desire to make a statement about the state of the nation.
In notes left behind, Livelsberger denied wanting to harm anyone but himself, writing that the explosion was intended as a “wake-up call” for the American public. He described feeling burdened by the lives lost during his service and the toll of combat.
In a statement, OpenAI expressed sadness over the incident and reiterated its commitment to responsible AI use. The company emphasized that its models are designed to discourage harmful instructions and that the information provided by ChatGPT was publicly available online. OpenAI is cooperating with law enforcement in the ongoing investigation.
The incident underscores the growing challenge of balancing technological advancement with safety. Law enforcement agencies nationwide are now examining how generative AI tools like ChatGPT could be misused in criminal activities, marking a critical moment for policy and prevention strategies.
