A 25-year-old man detonated a car bomb outside a Palm Springs fertility clinic in May 2025, killing himself and injuring four others in an attack that federal law enforcement officials later tied to unmonitored artificial intelligence prompts. The suspect used an unnamed generative AI tool that failed to detect his request for bomb-making instructions, according to a previously unreported law enforcement document reviewed by The Guardian.
Federal Warnings Over Unchecked Bomb-Making Instructions
The privacy features of the suspect’s chat sessions hindered police from receiving early alerts that someone was researching explosive devices, according to the bureau’s Los Angeles division memo distributed in May 2026. FBI officers warned in the document that threat actors will likely increase their use of generative AI tools for attack planning because current built-in safeguards remain insufficient.
Malicious actors continue to successfully employ jailbreaking techniques to access dangerous information, the memo stated. While AI executives have frequently focused their public remarks on long-term, hypothetical doomsday scenarios, law enforcement agencies have spent years grappling with the immediate violent crimes enabled by the technology.
Intelligence Memos Track Anti-Tech Extremism and Datacenter Threats
A cache of more than 80 law enforcement and intelligence agency documents distributed between 2021 and 2026—obtained by the government transparency non-profit Property of the People and shared with The Guardian—shows that agencies including the Department of Homeland Security, the FBI, and the New York City Police Department have monitored artificial intelligence as both a tool and a target for extremists.
Before the widespread public rollout of ChatGPT, far-right militia members and domestic violent extremists targeted critical infrastructure to instigate social collapse. Seth Pendley, a January 6 rioter, plotted to blow up an Amazon Web Services facility to disrupt 70% of internet service before pleading guilty and receiving a 10-year prison sentence in 2021.
As the generative AI boom accelerated, fusion centers across the United States reported a surge in online rhetoric vowing to destroy datacenters. A 2025 intelligence bulletin from the Philadelphia police department cited a white supremacist social media account threatening to use a mixture of gasoline and Tannerite against an AI facility.
Surveillance Concerns Surrounding Mainstream Opposition to AI Infrastructure
Law enforcement tracking has increasingly overlapped with mainstream public dissent, raising questions among civil liberties advocates. Documents from the Nevada Threat Analysis Center in April 2026 highlighted local city council hearings and zoning moratorium debates regarding water consumption, utility costs, and pollution from upcoming datacenters as part of a rising wave of public opposition.
Ryan Shapiro, executive director of Property of the People, criticized the broad intelligence gathering. There is nothing extreme about not wanting to lose a livelihood, drinking water, or privacy to runaway artificial intelligence, Shapiro said, arguing that treating legitimate public objections as security threats is what remains extreme.
Common Questions Regarding Artificial Intelligence and Security
Did law enforcement track anti-datacenter threats before generative AI tools became popular?
Federal bulletins show agencies monitored far-right militia plots against critical infrastructure as early as 2021, long before the late-2022 public launch of ChatGPT. Convicted bomb plotter Seth Pendley targeted an Amazon Web Services facility in a failed 2021 attack aimed at disrupting internet service.
How do bad actors bypass AI safety guardrails for attacks?
Law enforcement memos state that malicious individuals utilize jailbreaking techniques to circumvent built-in restrictions on generative AI chatbots, allowing them to extract bomb-making instructions and tactical attack guidance without triggering automated police alerts.
What specific security risks do police face from consumer technology?
New York and Maine intelligence bulletins warned that police officers face harassment and operational risks from AI-generated fake evidence, deepfakes, and smart-doorbell software features that automatically notify residents of nearby police presence.