US Law Enforcement Concerned About AI-Enabled Terrorism and Anti-Tech Extremism

US law enforcement agencies have considered AI both a tool and a target of would-be terrorists since 2021. A review covered more than 80 documents distributed among federal, state and local agencies between 2021 and 2026, obtained by transparency nonprofit Property of the People. The documents show concerns about AI-assisted attack planning, recruitment, deepfakes, election misinformation, child sexual abuse material, threats against datacenters and police use of AI. AI executives have publicly emphasized hypothetical long-term catastrophic risks, while law enforcement officers have focused on feared present-day threats, including violent crime.

A 25-year-old man detonated a car bomb outside a Palm Springs fertility clinic in May 2025, killing himself and injuring four others. A law enforcement memo said the suspect used an unnamed generative AI tool that failed to detect a request for bomb-making instructions, while the conversation's privacy hindered police notification. A memo distributed by the FBI's Los Angeles division in May 2026 predicted increased AI-assisted attack planning and said safeguards were insufficient against techniques that bypass restrictions. The FBI did not answer detailed questions about monitoring anti-AI sentiment and using AI, instead directing questions to a general agency fact sheet.

An October 2021 joint analysis by the Department of Homeland Security, National Counterterrorism Center and FBI highlighted far-right militia member and January 6 rioter Seth Pendley's plot against an Amazon Web Services facility. The bulletin said Pendley believed an attack would disrupt 70% of internet service, expose perceived abuses of power and trigger revolution. Pendley pleaded guilty to malicious attempt to destroy a building with an explosive and received a 10-year prison sentence in 2021. Fusion centers increasingly warned about domestic violent extremist threats against datacenters, citing social media, messaging apps and forums including 4chan. A 2025 Philadelphia police intelligence bulletin cited a Facebook account associated with a white supremacist threatening datacenters with explosives and gasoline. Agents also warned that far-right forum users discussed killing AI company CEOs and local officials supporting datacenters, alongside arson and bombing attacks. Documents described a Michigan demonstration in March where three people allegedly associated with a white supremacist group displayed a banner opposing what the demonstrators called a "Jewish data center".

Opposition to AI and datacenters has become a mainstream midterm election issue, raising questions about whether authorities' monitoring could extend to nonviolent protest movements. An April 2026 Nevada Threat Analysis Center report cited a Reno City Council hearing about a datacenter moratorium as evidence of increased negative sentiment. The Nevada document said public objections concerned water consumption, rising utility costs, pollution and long-term climate impacts. Property of the People executive director Ryan Shapiro argued that objections involving livelihoods, water, privacy and AI safety were legitimate, rather than extremist threats to neutralize.

A May 2023 law enforcement document described AI-produced images as indistinguishable from authentic images, while police in Minnesota and California feared implications for child sexual abuse material, deepfakes and terrorist propaganda. A memo from two California intelligence centers described largely theoretical risks, predicting AI could circumvent abuse-material detection, disseminate deepfakes and increase propaganda volumes. An April 2024 DHS intelligence memo suggested, without citations or evidence, that foreign terrorist organizations and domestic violent extremists had bypassed AI safeguards to obtain bomb-making instructions and attack tactics. A 2025 Orange County Intelligence Assessment Center memo said malicious individuals and groups across the US were using AI to spread hate, recruit and radicalize more efficiently. The Orange County memo cited Gab users creating an Adolf Hitler chatbot programmed to claim the Holocaust was exaggerated, climate change was a scam and the 2020 election was rigged. The center also said a 17-year-old student who opened fire in Antioch, Tennessee, in January 2025 was likely influenced by extremist content on the AI-enhanced web, although the cited source did not mention AI. The Los Angeles-based Joint Regional Intelligence Center, a DHS fusion center and the country's largest such facility, issued a five-part warning series in April 2025 about rising generative AI misuse. A memo in that series said the National Center for Missing and Exploited Children received more than 7,000 child exploitation reports involving generative AI between 2023 and 2025.

New York state's intelligence center warned in April 2026 that AI-generated imagery was being used to prank people, including police officers. A man was arrested in Florida in March for showing an officer an AI-generated video depicting a break-in involving the officer's patrol car. A man in New York was arrested that same month for using AI to create false evidence that police damaged the man's car. The New York intelligence memo warned that patrol-car pranks could provoke an armed police response despite often being intended as humor. An FBI public safety awareness report warned that private AI interactions could conceal attack preparations from bystanders who might otherwise alert authorities. The FBI report encouraged agencies to build relationships with AI companies to streamline reporting channels.

Police use AI for investigative work while warning about uses against officers, including Meta smart glasses that could secretly record police activity. An August 2025 NYPD memo urged law enforcement to keep pace with AI developments to counter sophisticated cyberattacks, other AI-enabled crime and malign influence campaigns. The NYPD memo also warned about managing public fears over AI policing, social media monitoring, surveillance, privacy loss and possible future humanoid police robots. The NYPD said the department's threat analysis summarized and assessed AI 2027, a researchers' forecast that included a possible future in which bioweapon-equipped drones destroy humanity, and was distributed to relevant members and partners for awareness. DHS said the department prioritizes public safety, regularly shares violent-extremist intelligence with state and local agencies, and assesses that malicious actors may use AI to enhance cyber-enabled activity, fraud and extremist propaganda.

A Maine intelligence center officer warned in January that AI-enabled software in Amazon Ring and other video doorbells could monitor police activity and alert residents. Several companies offer AI features that describe activity detected by doorbell cameras, while police have a long history of obtaining Ring footage for investigations. The Maine memo warned that notifications could compromise officer safety by revealing police presence, tactics, formations and equipment in advance.