On January 1, 2025, a deadly explosion involving a Tesla Cybertruck outside the Trump Hotel in Las Vegas left one person dead and seven others injured.
Investigators have now revealed that the attacker, Matthew Livelsberger, 37, from Colorado, used the AI platform ChatGPT to help plan the attack.
According to police, Livelsberger’s actions were detailed in a six-page manifesto, and evidence shows he utilized ChatGPT to gather information about explosives.
He sought advice on how to conduct the plot, asking questions like “Purchasing Tannerite in Arizona how much do you buy?” and “What pistol could set it off?”
Sheriff Kevin McMahill confirmed the findings in a press conference on January 8, noting that Livelsberger had used the AI tool to determine how much explosive material would be needed and other aspects of the attack.
The investigation revealed that Livelsberger believed he was being followed and documented his mental health struggles, including past traumatic military experiences.
He also expressed feeling like a “shell of a human being” with “nothing to live for,” suggesting the attack had a suicidal component.
ChatGPT’s developers, OpenAI, acknowledged the incident, emphasizing their commitment to ensuring AI tools are used responsibly.
They confirmed that ChatGPT had provided publicly available information and cautioned against harmful or illegal actions, though Livelsberger’s questions still garnered the information he sought.
The Las Vegas Metropolitan Police Department is continuing its investigation but has withheld certain details due to the potential relevance to national security.
