Authorities said 37-year-old Matthew Alan Livelsberger used the artificial intelligence platform in the crime through ChatGPT to obtain information on how to carry out an explosion.
The former soldier, who blew up a Cybertruck in front of the entrance to the Trump International Hotel in Las Vegas on New Year’s Day, is believed to have relied on artificial intelligence for guidance on how to execute the blast, according to statements from officials on Tuesday.
Authorities stated that Livelsberger asked ChatGPT how to assemble an explosive device, the speed required to ignite the explosives inside the truck rather than just having them burn, and what laws need to be bypassed to obtain the necessary materials.
Read also: The growing benefits and risks of artificial intelligence: An urgent call for comprehensive regulation
Las Vegas Police Sheriff Kevin McMahill said, “We know that artificial intelligence is going to change the game for everyone in one way or another in our lifetimes. And I believe this is the first incident that I am aware of on U.S. soil where ChatGPT was used to help someone build a specific device.”
Police identified Livelsberger last week as the primary suspect in the explosion. According to the Clark County Coroner’s Office, he died from a self-inflicted gunshot wound to the head, and his severely burned body was identified through familial DNA and tattoos.
Authorities showed video footage depicting Livelsberger pulling a racing fuel can from the back of the Cybertruck and pouring fuel onto the area containing the explosives.
Kenny Cooper, assistant special agent in charge of the San Francisco field division of the Bureau of Alcohol, Tobacco, Firearms and Explosives, said that the gunshot Livelsberger fired at himself may have triggered the explosion due to the ignition of fireworks and other explosive materials inside the truck.
Investigations indicated that no evidence was found suggesting the use of a remote detonation system or a complex timing system.
A spokesperson for OpenAI, which owns ChatGPT, said: “We are saddened by this incident and are committed to ensuring AI tools are used responsibly. Our models are designed to refuse harmful instructions and minimize harmful content. In this case, ChatGPT provided information already publicly available on the internet along with warnings against harmful or illegal activities.” The spokesperson added that the company is cooperating with authorities in the investigation.
Law enforcement concerns have grown regarding the use of artificial intelligence as a guidebook for crimes, with a special report from NBC News highlighting the challenges associated with artificial intelligence. Rebecca Weiner, the New York Police Department’s deputy commissioner for intelligence and counterterrorism, said: “AI amplifies existing problems and makes them more widespread and at a lower cost. Technology makes it easier to do troubling things.”
Despite this, law enforcement agencies in Las Vegas and New York still lack sufficient cooperation from AI services to receive alerts when queries related to carrying out attacks or building explosives are made.
Last week, authorities revealed that an application on one of Livelsberger’s phones displayed messages criticizing the U.S. government and admitting that he intentionally blew up the vehicle.
The messages included a letter addressed to “soldiers, peers, and the American people,” calling on them to “wake up” due to the country’s poor leadership. In another message, he claimed that the explosion was “not a terrorist attack” and that Americans only pay attention to “violent spectacles.”
Authorities also found a six-page manifesto on his phone containing grievances about the current administration and military experiments that he described as repeatedly haunting him.
