Eight years after Edward Snowden revealed the NSA’s mass data collection infrastructure, that infrastructure has been weaponized. The passive dragnet has been transformed into an active, AI-powered targeting system capable of tracking millions of people in real time, predicting “threats” before they occur, and automating the machinery of deportation, surveillance, and political repression.
What follows is not speculation. It is documented fact.
In April 2025, US Immigration and Customs Enforcement granted a $30 million contract to Palantir Technologies—the data-analytics firm founded by Peter Thiel—to build “ImmigrationOS,” a surveillance platform designed to “streamline the identification and apprehension of individuals prioritized for removal” . The contract justification explicitly cited support for Presidential Executive Orders, including those targeting visa overstays and “violent criminals” .
But the platform does far more than track undocumented immigrants.
ImmigrationOS sits on top of Palantir’s existing data-integration stack—technology originally built for military and intelligence work and later adapted for domestic law enforcement . These platforms ingest structured and unstructured data from dozens of sources: criminal records, civil immigration files, state motor vehicle department records, Social Security Administration data, local law enforcement databases, jail and court systems, and commercial aggregators that assemble information on utilities, phones, and financial activity .
The system then normalizes identifiers such as names and addresses and exposes them through search and analytic tools that generate what ICE agents call “targeting packages” .
According to internal documents obtained by 404 Media, Palantir is developing a specific tool for ICE called ELITE—Enhanced Leads Identification & Targeting for Enforcement. ELITE maps potential deportation targets, generates individuals’ dossiers, and assigns a “confidence score” to a person’s current address, obtained from HHS and other federal sources . The user guide for ELITE, later published by 404 Media, reveals that during “special operations”—which target “groups of pre-defined aliens”—normal safeguards in the system may need to be turned off .
Let that sink in. The safeguards are optional.
In October 2025, The Lever uncovered that ICE is paying $5.7 million to use an AI-powered social media monitoring platform called Zignal Labs . The platform is capable of analyzing more than 8 billion posts per day in over 100 languages, using machine learning, computer vision, and optical character recognition to process data into “curated detection feeds” that ICE can use to flag individuals for deportation .
Zignal Labs captures geolocated images and videos. It identifies emblems and patches to “confirm the operators involved” in activities. It notifies “operators on the ground” in real time .
This means ICE can trace someone’s location based on a TikTok video, a Facebook photo, or an Instagram story. The agency can track your movements, your associations, your political activities—all without a warrant, all without your knowledge.
But Zignal Labs is just the beginning.
WIRED reports that ICE plans to hire nearly 30 contractors to comb through content on Facebook, Instagram, TikTok, X, YouTube, and other platforms 24/7, with some required to be available “at all times” . The plan calls for continuous monitoring with defined turnaround times, and seeks to incorporate advanced surveillance technologies from companies including ShadowDragon, Clearview AI, and Palantir .
A document seen by WIRED reveals that these analysts may even search for data about a target’s family members, friends, or coworkers to pinpoint their whereabouts for ICE officers .
David Greene, civil liberties director at the Electronic Frontier Foundation, puts it plainly: “Automated and AI-powered monitoring tools will give the government the ability to monitor social media for viewpoints it doesn’t like on a scale that was never possible with human review alone” .
The chilling effect is the point. As Will Owen of the Surveillance Technology Oversight Project testified, “The scale of this spying is matched by an equally massive chilling effect on free speech” .
The government cannot legally wiretap you without a warrant. So instead, it buys the data.
In April 2025, a congressional subcommittee heard testimony from civil libertarians and conservative legal scholars about what they called the “data broker loophole” . By purchasing information from private data brokers—location history, search records, financial transactions, app usage—the government sidesteps Fourth Amendment protections that would otherwise require warrants .
James Czerniawski, a senior policy analyst for Americans for Prosperity, testified that “the purchase of such information by the government creates an additional threat, as law enforcement can then use this data tied to Americans engaging in constitutionally protected activities and subject them to additional surveillance via other technologies” .
The Office of the Director of National Intelligence recently released a partially declassified report admitting that the Intelligence Community is “collecting increasing amounts of commercially available information” but does not know how much it is collecting, what types, or what it is doing with the data .
They don’t know. And they don’t have to tell you.
In Cook County, Illinois—a sanctuary jurisdiction that limits cooperation with civil immigration enforcement—jail records can still reach ICE after being sold to companies like LexisNexis Risk Solutions, which then licenses access back to federal agencies . Local officials may not even realize how their data travels once it enters the commercial market .
The contracts are designed to be opaque. The data flow is designed to be deniable. The surveillance is designed to be invisible.
The Department of Homeland Security has deployed facial recognition at all US international airports and seaports . The FBI, US Marshals, and the Child Exploitation and Obscenity Section of DOJ are among the primary operators of facial recognition technology . The Bureau of Alcohol, Tobacco, Firearms and Explosives conducted at least 549 searches using commercial facial recognition systems like Clearview AI and Vigilant Solutions over a two-and-a-half-year period .
Clearview AI’s database includes tens of billions of images scraped from the public internet . ICE holds a facial-recognition contract with the company, using it to search for matches on “priority suspects” .
A February 2025 Government Accountability Office audit revealed that ATF lacked any dedicated policy governing the use of these tools, had not conducted a risk assessment, and failed to require staff training on the proper use and limitations of facial recognition technology .
Despite ATF’s claim to have halted its use of commercial services by April 2023, a Senate report on the July 2024 attempted assassination of President Trump revealed that ATF was apparently requesting photos of the shooter for facial recognition purposes as late as August 2024 .
Lawmakers are now concerned that ATF may be circumventing restrictions by relying on state and local agencies to conduct searches on its behalf—an indirect use that does not absolve the agency of its constitutional obligations .
The US Commission on Civil Rights warned in a 194-page report that these tools are being used in a “regulatory vacuum” that lacks adequate standards, oversight, and training . The commission cited troubling disparities in the accuracy of facial recognition systems across race and gender lines, opening the door to discriminatory outcomes .
Commissioner Mondaire Jones underscored how disparities in accuracy, insufficient oversight, and lack of access to legal recourse create barriers to justice—particularly for vulnerable populations .
In March 2025, Axios reported that the State Department had launched a new social media surveillance program called “Catch and Revoke” . The program uses artificial intelligence to assist in reviewing “tens of thousands of student visa holders’ social media footprints” to find “evidence of alleged terrorist sympathies expressed after Hamas’ attack on Israel” .
Posts are taken out of context. Speech is misinterpreted by algorithms. And visas are revoked without due process.
The Bulletin of the Atomic Scientists describes this dynamic precisely: “AI’s promise of behavior prediction and control fuels a vicious cycle of surveillance which inevitably triggers abuses of power” .
The same month, the State Department revoked the visas of six people who the US claimed “celebrated” the shooting of right-wing commentator Charlie Kirk . Days later, ICE arrested nine street vendors on New York City’s Canal Street—shortly after a conservative influencer tagged ICE in a post showing vendors in the area .
The pattern is unmistakable. The government is using AI-powered surveillance to target political dissent, to chill protected speech, and to punish those whose views deviate from the administration’s line.
Sacha Haworth, executive director of the Tech Oversight Project, told The Verge: “This is another example of Big Tech CEOs partnering with an increasingly authoritarian federal government as part of Trump’s ongoing attempts to clamp down on free speech. This should terrify and anger every American” .
Perhaps the most terrifying aspect of this surveillance architecture is that no one is watching the watchers.
The NSA’s XKeyscore program—first revealed by Snowden in 2013—continues to operate with no judicial oversight and limited congressional oversight, despite its potential to capture Americans’ communications . The program relies heavily on the “autonomous collection of massive data sets” and analysis driven by artificial intelligence .
Travis LeBlanc, a Democratic member of the Privacy and Civil Liberties Oversight Board appointed by President Trump, voted against approving the panel’s classified report on XKeyscore, saying that the board “failed to adequately investigate or evaluate” the NSA’s collection activities . He noted that the program resulted in hundreds of compliance incidents in 2019, a majority of which were considered “questionable intelligence activities”—meaning the action may have involved improper surveillance of Americans’ communications .
When asked about the NSA’s legal analysis, LeBlanc said it “lacks any consideration of recent relevant” privacy case law, including Supreme Court decisions imposing stricter limits on cellphone and geolocation surveillance .
The indexing work on the data is done by machines, before a human ever sees it. As Ashkan Soltani, a senior fellow at Georgetown University’s Institute for Technology Law and Policy, explained: “The realities of the internet today means the likelihood that the NSA might accidentally be processing an American’s communications is quite high” .
LeBlanc said the board did not even analyze the extent to which XKeyscore’s use of machine analysis—as opposed to human review—of Americans’ information triggers Fourth Amendment scrutiny .
They don’t know. They didn’t ask. And they won’t tell us.
DHS has awarded more than $1 billion in IT contracts over roughly the first year of President Trump’s second term, including tens of millions earmarked for Palantir’s work on ImmigrationOS . ICE is looking to spend more than $300 million for social media monitoring tools, facial recognition software, license plate readers, and services to find where people live and work .
The agency has tapped into license plate-scanning security cameras and gained access to tools that track the movement of millions of phones . It can request footage from more than 2,000 local police and fire departments that partner with Amazon’s Ring . Homeland Security Investigations uses specialized vendor tools to identify and track people of interest .
Another vendor, under a confidential agreement, provides object-matching capabilities—identifying recurring cars or clothing across multiple videos by tracking distinctive visual features such as damaged bodywork or tears in fabric .
This is not surveillance. This is omnipresence.
When citizens tried to fight back, the tech platforms complied with the administration’s demands.
Apps like ICEBlock, Red Dot, and DEICER allowed users to pinpoint where ICE agents were active, forming an online whisper network to alert potential targets . Eyes Up provided a way for users to record and upload footage of abusive law-enforcement activity, building an archive of potential evidence .
All were removed from Apple and Google’s app stores .
The Attorney General, Pam Bondi, explicitly targeted ICEBlock in a statement to Fox News Digital, claiming it “put ICE agents at risk just for doing their jobs” .
The creator of Eyes Up, a software developer who uses only his first name to separate the project from his professional work, described a “double standard”: applications of technology friendly to the administration’s goals go unchallenged, while tools that empower citizens to document abuse are suppressed .
“It’s clear whose rules they’re following, who they are trying to win over,” he said .
Bruce Schneier, renowned security expert and Harvard fellow, put it bluntly in January 2026: We have entered the era of “bulk spying” .
“AI is able to do speech-to-text and generate summaries,” Schneier said. “We are entering a world where we have mass surveillance plus bulk spying. I can guarantee that the US, China, Russia, and other countries are doing this” .
He warned that the surveillance state is repeating the mistakes of the social media era—but worse. “All the terribleness of social media is being replicated in AI in worse ways” .
Congressman Warren Davidson, an Ohio Republican, wrote in September 2025: “Surveillance capitalism and government spying on its own citizens has run amok. At the intersection, Palantir has been commissioned to develop an AI tool to unify the data in every federal database, turning it into useful, easily accessible information” .
He continued: “The Patriot Act massively expanded domestic surveillance. The Bank Secrecy Act ended any claim to privacy in your financial dealings. Most Bureaus of Motor Vehicles monetize the data that citizens are required to provide to drive or obtain a Real ID. Now the government is buying data that would otherwise require a warrant or subpoena—circumventing the Fourth Amendment” .
The Bulletin of the Atomic Scientists—the same organization that created the Doomsday Clock—now compares AI to nuclear technology: “radioactive, dangerous, and, yes, sometimes useful” .
The only way to prevent catastrophe, the Bulletin argues, is “restrictive, well-designed controls to prevent damage to democracy, much like humans did with nuclear energy” .
But those controls do not exist. Congress has not acted. The courts are moving slowly. And the administration is moving fast.
In the meantime, the machine learns. It watches. It predicts. It targets.
The data Snowden revealed is no longer sitting in databases, collecting dust. It is being fed into algorithms that decide who gets detained, who gets deported, who gets flagged as a “threat.” The surveillance state has been weaponized, and the weapon is AI.
This is not hyperbole. This is not speculation. This is documented fact, confirmed by government contracts, congressional testimony, investigative journalism, and the admissions of the agencies themselves.
The dystopia is here. The only question that remains is whether we will do anything about it before it’s too late.