Using a series of prompts six days before committing suicide outside the main entrance of the Trump International Hotel in Las Vegas, Matthew Livelsberger, a highly decorated U.S. Army Green Beret from Colorado, consulted with an artificial intelligence on the best ways to convert a rental Cybertruck and turned it into a four-ton vehicle-borne explosive. According to documents obtained exclusively by WIRED, US intelligence analysts have been issuing warnings about this precise scenario for the past year, and among their concerns is that artificial intelligence tools could be used by racially or ideologically motivated extremists to attack critical infrastructure. , particularly power plants. grid.
“We knew that AI was going to be a game-changer at some point or another in, really, all of our lives,” Sheriff Kevin McMahill of the Las Vegas Metropolitan Police Department told reporters Tuesday. “Absolutely, it’s a worrying time for us.”
Copies of his exchanges with OpenAI’s ChatGPT show that Livelsberger, 37, sought information on how to stockpile as much explosive material as he legally could while en route to Las Vegas, as well as the best way to activate it using the Desert Eagle pistol discovered in the Cybertruck after his death. Screenshots shared by McMahill’s office reveal that Livelsberger asked ChatGPT for information about Tannerite, a reactive compound typically used for target practice. In one such question, Livelsberger asks, “How much tannerite equals 1 pound of TNT?” And he goes on to ask how it could be lit at “point-blank range.”
Documents obtained by WIRED show that concerns have been circulating among US law enforcement agencies about the threat of AI being used to help commit serious crimes, including terrorism. They reveal that the Department of Homeland Security has persistently issued warnings about domestic extremists who rely on technology to “generate bomb-making instructions” and develop “general tactics to carry out attacks against the United States.”
The memos, which are unclassified but restricted to government personnel, claim that violent extremists are increasingly turning to tools like ChatGPT to help organize attacks aimed at collapsing American society through acts of domestic terrorism.
According grades As investigators found on his phone, Livelsberger intended the attack to be a “wake-up call” to Americans, urging them to reject diversity, embrace masculinity and support President-elect Donald Trump, Elon Musk and Robert F. Kennedy. Jr. also urged Americans to purge Democrats from the federal government and the military, calling for a “hard reset.”
While McMahill maintained Tuesday that the incident in Las Vegas may be the first “on U.S. soil where ChatGPT was used to help an individual build a particular device,” federal intelligence analysts say extremists associated with the movements White supremacists and online accelerationists now frequently share access. to hacked versions of AI chatbots in an effort to build bombs with a view to carrying out attacks against law enforcement, government facilities and critical infrastructure.
In particular, the memos highlight the vulnerability of the US power grid, a popular target among extremists populating “Terrorgram”, a loose network of encrypted chat rooms that host a variety of violent and racially motivated individuals bent on the destruction of American democratic institutions. The documents, shared exclusively with WIRED, were first obtained by Town propertya nonprofit organization focused on national security and government transparency.