The Tech Behind ICE’s Deportation Crackdown

The government’s drive to speed up removals is underpinned less by need-for-speed door-to-door sweeps than by a complex stack of surveillance, analytics and digital forensics. At the center of it all is Immigration and Customs Enforcement, which has used data from apps, phones and, most controversially, public records to track down targets for costliest raids. Public records, agency documents and reporting from outlets including CNN, 404 Media, Forbes, The Intercept and Wired reveal an extensive toolkit cobbled together from many vendors.

How ICE Picks Its Targets for Arrests and Deportations

Two arms drive the work: Enforcement and Removal Operations, which homes in on arrests and deportations; and Homeland Security Investigations, which fulfills prosecution of more complex criminal investigations that channel data back into enforcement operations. At its heart, the playbook combines three components: mass data collection; robust search and scoring development tools to filter that data down into usable intelligence; and field tech — literally the fingerprint reader itself — turning intel into arrests.

Phone Tracking and Cell-Site Simulators Used by ICE

ICE uses cell-site simulators — also known as stingrays or IMSI catchers — that trick nearby phones into connecting with them and revealing sensitive information. That allows agents to identify and locate devices, and in some cases extract metadata. According to procurement records, ICE has paid TechOps Specialty Vehicles more than $1.5 million for custom-designed surveillance vehicles, including platforms to hold cell-site simulator gear. Civil liberties groups have long criticized this approach, because it nabs the phones of bystanders and courts faulted the secrecy around deployments; in one Baltimore case, prosecutors dropped charges rather than disclose use of the technology.

Facial Recognition at Scale Across Massive Databases

ICE is an agency contractor to Clearview AI, which has built a gargantuan database of images scraped from the public web that can match photos with billions of others on the service. Government procurement databases show contracts worth several million dollars, including enterprise licenses and forensic software. Clearview is framing the tool as a way to help identify suspects and victims in cases of online exploitation; skeptics have highlighted documented accuracy gaps and demographic bias from the National Institute of Standards and Technology. Even when employed for serious offenses, these systems can bleed into immigration enforcement with shared leads and databases.

Spyware and Device Exploits for Unlocking Phones

A second line of activity has been with spyware and phone unlocking tools. ICE has contracted with Paragon Solutions for their own proprietary system which is a package deal that includes licenses, hardware, maintenance and training. But the deal raised questions under a federal policy intended to limit government use of commercial spyware associated with human-rights risks. And agents elsewhere rely on various tools, including Magnet Forensics software and GrayKey hardware — now under one company after a recent round of acquisitions — to unlock iPhones as well as Android devices, with the most recent contract estimated at about $3 million for Magnet. Device exploitation condenses investigative legs by extracting texts, chat logs and app data that can situate a person at a location or chart connections.

Buying Ad Tech Trails and Commercial Location Data

In addition to direct tracking, ICE also purchases access to commercial caches of historical location data. Reporting by 404 Media and Forbes reveals that the agency has been employing Penlink tools — Webloc and Tangles — to search billions of mobile location pings against social media and other open sources. These sets of data are extracted from app software kits and real-time bidding in the digital advertising market, and resold by data brokers. Because the information is bought, agencies have sometimes argued that they can access it without a warrant, an argument privacy advocates say skirts around the Supreme Court’s 2018 ruling in Carpenter. The legal gray zone here is illustrated by the fact that while both location brokers and some of Congress’s proposals, such as the Fourth Amendment Is Not For Sale Act, fall under scrutiny from the Federal Trade Commission.

Data Brokers and Risk Scoring Inside ICE Operations

ICE also relies on the publicly available court records to compile dossiers. Documents obtained under the Freedom of Information Act revealed over a million searches in a matter of months using LexisNexis’s Accurint Virtual Crime Center, which connects utilities, addresses, vehicle records and associates. According to The Intercept, the system is utilized in order to alert bureaucrats when it finds “suspicious” behavior and proactively investigate migrants. Today, ICE already spends millions on a subscription to a law enforcement investigative database from LexisNexis, which the vendor says is subject to compliance and public safety obligations.

Palantir’s Data Spine for Immigration Enforcement

Palantir’s Investigative Case Management — a database, search case management tool and “end-to-end collaboration platform” for federal law enforcement that stitches together dozens of databases (immigration records, criminal history, biometrics), tips and travel data through machine learning analytics with little human involvement — sits right in the center of this stack. Contracts range from an $18.5 million extension to its parent, another agreement north of $90 million. An earlier examination of ICM by 404 Media refers to granular filters for visa types, ports of entry and even physical traits, affiliations and location histories. Palantir also has a project called ImmigrationOS, as Business Insider and Wired have reported, which boasts end-to-end orchestration — from triaging leads to tracking overstays and even tallying up self-deportations in near real time.

What It Costs, and Why This Technology Matters

Do the math on just a few of those recent purchases — for cell-site simulator vehicles, facial recognition software, spyware to track mobile devices, ad-technology location tools or data brokers and Palantir — and public records show tens of millions of dollars in current spending. CNN has reported hundreds of thousands of deportations within a small window, with ICE representing a significant portion — demonstrating that this technical infrastructure is not merely theoretical but in use.

Risks of Oversight and the Question of Accuracy

Failures and correct results are both amplified by scale. Open-source location data can be dirty; face recognition can make mistakes; stingrays hoover up bystanders; and risk scores carry bias from the underlying data. The Department of Homeland Security’s Privacy Office and Office for Civil Rights and Civil Liberties have issued impact assessments and guidelines, the DHS Inspector General has sounded alarm bells about governance gaps in data programs. But when commercial data is combined with powerful analytics, the result can outstrip policy at a rate that calls into question longstanding protections around warrants, minimization and retention periods of information on non-targets.

The upshot: ICE is cracking down using an integrated data-first model. It combines commercial surveillance with government databases, and produces leads to send agents equipped with phone unlockers and mobile tracking devices. Whether legislators and courts redraw the lines that define data buying, facial recognition and phone tracking will determine how far this model can travel — and how many people get caught up in its net.