4 min read
Recent attacks in Las Vegas and New Orleans show how AI tools are being used by terrorists.
The attacks—one involving ChatGPT's language and reasoning capabilities, the other exploiting Meta's smart glasses technology—are raising questions about how to strengthen protective measures while preserving AI's beneficial applications.
Authorities say that U.S. Army Special Forces Master Sgt. Matthew Livelsberger—a decorated Green Beret who had shown signs of PTSD—methodically employed ChatGPT to research crucial elements of his Tesla Cybertruck attack outside Trump International Hotel in Las Vegas on January 1.
According to Las Vegas Sheriff Kevin McMahill, Livelsberger's interaction with the AI platform extended far beyond casual queries. The 37-year-old soldier used ChatGPT to calculate precise explosive quantities, research firework procurement channels, and explore methods for purchasing phones without leaving digital traces.
"We knew that AI was going to change the game at some point or another in really all of our lives," McMahill said. "This is the first incident that I'm aware of on U.S. soil where ChatGPT is utilized to help an individual build a particular device."
The incident culminated in Livelsberger pouring racing fuel over the rented Cybertruck before taking his own life, with investigators suggesting the muzzle flash from his firearm may have triggered the explosion. Seven people were injured, though the hotel structure remained largely undamaged and the incident didn’t claim any lives beyond the perpetrator's own.
The same day as the first incident, 42-year-old U.S. Army veteran Shamsud-Din Jabbar drove a rented pickup truck into a crowd during the new year celebration.
Prior the incident, Jabbar used Meta's Ray-Ban smart glasses, powered by Meta AI, to conduct detailed reconnaissance of his target area.
FBI investigations revealed that Jabbar made two visits to New Orleans' French Quarter—on October 30 and November 10—during which he used the smart glasses to capture extensive footage of Bourbon Street. The glasses allowed him to record video and photos hands-free while cycling through the area. FBI Special Agent Lyonel Myrthil detailed how Jabbar methodically recorded the street layout, repeatedly scanning left to right to capture comprehensive viewpoints.
The FBI published one of the videos recorded by Jabbar:
The attack itself claimed 14 lives and left 35 injured, including two police officers, when Jabbar drove a pickup truck through New Year's crowds at 3am. Though he wore the Meta glasses during the attack, he did not activate them for recording or livestreaming.
Recovered footage showed Jabbar testing the glasses' capabilities, examining his reflection and ensuring optimal recording angles. Counterterrorism expert Sam Hunter emphasized the tactical advantage such technology provided.
“You’re really getting a sense of the eyeline and eyesight, and all the things that you’re going to want to look out for if you’re trying to plan an attack,” Hunter said in a statement shared with NBC. “I would not be surprised if you see versions of them or folks using them for attack planning in the future, again because they’re so discreet in terms of capturing that footage,”
Tech companies and regulators have been working to strengthen safeguards while preserving AI's beneficial applications.
OpenAI emphasized that ChatGPT incorporates specific protections against harmful instructions, and provides explicit warnings about illegal activities. The company said that it has fully cooperated with law enforcement investigations, while simultaneously working to enhance its safety protocols.
“Real-world use has also led us to develop increasingly nuanced policies against behavior that represents a genuine risk to people, while still allowing for the many beneficial uses of our technology,” OpenAI said in an official blog post.
That said, as Decrypt has reported, simple jailbreaking techniques have proven effective when applied against these models: Basic methods like writing in leetspeak or using continuous roleplay can be enough to obtain illegal content and bypass such restrictions.
The company has not released a statement regarding the use of its tools by Sgt. Matthew Livelsberger for his terrorist attacks.
Meta, facing scrutiny over its smart glasses' potential for misuse ever since their release, has acknowledged the need to balance innovation with security concerns. The glasses' recording indicator light, while present, has been criticized as being too subtle.
Questioned about the terrorist attacks, a Meta spokesperson told NBC that the company is actively “in touch with law enforcement on this matter.”
Edited by Andrew Hayward
Decrypt-a-cookie
This website or its third-party tools use cookies. Cookie policy By clicking the accept button, you agree to the use of cookies.