An in-depth look at how ICE, under Obama and then Trump, has aggressively used big data, fed by social media, to clamp down on undocumented immigrants
a testament to the agency's quiet embrace of big data. https://www.nytimes.com/... “Surveillance works best when you don't notice it.” Tim Shorrock / @timothys : NYT reports that, via public record requests, it has “extensive proof that ICE relies on state D.M.V. databases and information products like CLEAR, from Thomson Reuters, to target immigrants.” Reuters, you might ask? Yes, you heard that right. https://www.nytimes.com/... ngel S. Daz / @angelsdiaz_ : A comprehensive account of the technologies that power the American deportation machine. I read this nodding along to all the connections being drawn while struggling to keep my heart from dropping into my stomach. https://www.nytimes.com/... Aaron Reichlin-Melnick / @reichlinmelnick : This article makes a compelling case that we're soon approaching the era of algorithmic immigration enforcement. This should trouble even the staunchest defenders of ICE—you need only look at China to see the direction that infrastructure leads toward. https://twitter.com/... April Glaser / @aprilaser : this made me cry. technology is not and has never been a neutral tool https://www.nytimes.com/... Peter Sterne / @petersterne : ICE increasingly resembles a secret police force that surveillance and then disappears ordinary people. And it acts in accordance with a parallel system of law (immigration courts) that lacks the traditional protections of our judicial system. That's scary. https://twitter.com/... NYT / @nytmag : The winter after Donald Trump was elected president, two strangers arrived in a parking lot on Washington State's Long Beach Peninsula. For weeks, they sat in their vehicle and watched the workers arrive in their trucks. Then the workers began to vanish https://www.nytimes.com/... McKenzie Funk / @mckenziefunk : ICE officers rolled in knowing names, nicknames, numbers, addresses, license plates, emails, social-media handles, patterns of life. “They knew everything,” residents said. I decided to find out how. Many months + many records requests later, some answers: https://www.nytimes.com/... McKenzie Funk / @mckenziefunk : Sixth lesson: You can't just delete your account, not with utilities and phone companies selling your info to data brokers, not with facial recognition and automated license plate readers spreading to every corner. We're datafying the physical space. McKenzie Funk / @mckenziefunk : Thread: In late 2017, I heard about an immigration crackdown in the supposed sanctuary state of Washington. ICE arrested a mom selling a piñata on Facebook. A grandpa who'd filed U.S taxes for almost 20 years. A school social worker. Dozens from one coastal county. McKenzie Funk / @mckenziefunk : Fourth lesson: Without social media and without private contractors providing data services, including @thomsonreuters, @VigilantSol, and many, many others, ICE would have a harder time finding people where they live. McKenzie Funk / @mckenziefunk : A second lesson: It's harder to stop data-sharing than it was to start. It's not that norms are changing and our personal info is just out there on social media, not just that. It's that after 9/11 we networked system after system so we'd never again fail to “connect the dots.” Diana Budds / @dianabudds : “What may be most unusual about Washington State is not what it collects and not what it has shared but the degree to which it has been forced to become transparent about the vast quantity of personal data that courses through its bureaucracy.” https://www.nytimes.com/... One Ring / @hypervisible : “If we can have completely accurate and completely pervasive surveillance, people won't notice and it won't be surveillance.” Yes. That's wrong. It's all wrong. pic.twitter.com/asJQDP38Ps William Fitzgerald / @william_fitz : this is the best story i've read about how ice is using tech. it's truly horrifying what we've created and enabled. must read for anyone interested in #notechforice https://www.nytimes.com/... David Beard / @dabeard : A piñata, a Facebook ad and poof: How ICE makes longtime residents disappear. Pictured: He lost his girlfriend, they lost their mother. https://www.nytimes.com/... https://twitter.com/... One Ring / @hypervisible : CLEAR, a Thomson Reuters collect it all tool “..isn't restricted by protections on what data the government can collect or keep — because it isn't government-owned.” 😐 https://twitter.com/... One Ring / @hypervisible : LP reader company has “...more than five billion historic images captured continually and automatically, thousands per minute, by infrared devices attached to lampposts and police cars and repossession-agent vehicles across the United States.” https://www.nytimes.com/... https://twitter.com/...
Context & Ripple Effects
McKenzie Funk's NYT investigation documents how ICE built its deportation apparatus less on raids than on data: state DMV databases, Thomson Reuters' CLEAR product, license-plate readers, facial recognition, and social media monitoring — with the Long Beach Peninsula arrests as the case study of what integrated lookups enable in the field.
What read in 2019 as an early warning has since become a documented pattern: Clearview AI's scraped-image facial recognition tool spread across hundreds of law enforcement agencies months later, Democratic senators prompted a DHS inspector general probe into buying brokered cellphone location data without warrants, and by 2023–2025 ICE was running GOST to score visa applicants' social media posts as derogatory and field-testing biometric identification apps. The story matters because it names the vendor layer — Thomson Reuters, Vigilant Solutions, and peers — that turns commercial data products into enforcement infrastructure.
First-order effects
- Undocumented immigrants in data-rich jurisdictions lose the practical anonymity that once separated daily life from enforcement risk: ICE can resolve a name into a home address, vehicle history, and family connections before agents leave the office.
- Commercial vendors like Thomson Reuters gain a durable government revenue stream for CLEAR-style lookups, while state DMVs become de facto arms of federal immigration enforcement through routine record-sharing.
Second-order effects
- Data brokers and surveillance vendors face a widening accountability surface — FOIA litigation, inspector general probes, and senatorial pressure — that forces them to defend or restructure warrantless data sales to DHS components.
- Rival agencies adopt the same playbook: State Department and DEA uptake of social-media surveillance tools like ShadowDragon shows ICE's procurement choices becoming a template across the enforcement portfolio.
Third-order effects
- If the pattern holds, immigration enforcement becomes algorithmic end-to-end — scoring, locating, and identifying people through fused commercial datasets — outpacing privacy law written for single-agency record checks.
- The structural fault line moves to the public-data permission boundary: whether scraping, brokerage purchases, and DMV sharing count as lawful collection when no court order exists, a question now being fought through oversight bodies rather than legislation.
The trend: Government enforcement is migrating from case-by-case investigation toward standing pipelines of commercially sourced data, with oversight bodies racing to decide which acquisitions require a warrant at all.