The Machine Knows Your Name

On a Tuesday morning in February 2024, ICE agents showed up at a home in Chicago. They didn't have a judicial warrant. They had something better: a target list generated by algorithms fed on utility records, DMV photographs, and commercial data purchased from brokers who track every American's movements — citizen or not.[1]

The $2.8 Billion Machine

Between 2008 and 2023, ICE spent at least $2.8 billion on surveillance technology, according to a Georgetown Law Center on Privacy & Technology investigation.[2] The agency has contracts with data brokers like LexisNexis, Palantir, and Thomson Reuters CLEAR — systems that aggregate billions of records from credit bureaus, phone companies, utility providers, and state DMVs into searchable profiles of virtually every adult in the United States.

None of this requires a warrant. ICE purchases the data on the open market, sidestepping Fourth Amendment protections that would apply if the government collected it directly. A 2022 ACLU analysis found that ICE had scanned the driver's license photos of one in three American adults through facial recognition queries to state DMV databases — without those people's knowledge or consent.[3]

“ICE now operates as a domestic surveillance agency that happens to do immigration enforcement. The data infrastructure they've built doesn't distinguish between citizens and non-citizens — it tracks everyone.”
— Nina Wang, Georgetown Law Center on Privacy & Technology

Palantir's Web

At the center of ICE's data operation sits Palantir's Investigative Case Management system, or ICM. Built under a contract worth over $200 million, ICM integrates data from more than a dozen federal databases — including the FBI, DEA, and the State Department — into a single interface that lets agents map relationships between people, places, phone numbers, and vehicles.[4]

In 2019, documents obtained through FOIA revealed that ICM was being used not just for deportation cases but to identify and track the sponsors of unaccompanied migrant children — the family members who come forward to care for kids in government custody. Within months of the program's expansion, sponsor referrals dropped sharply. Families stopped coming forward. Children stayed in detention longer.[5]

The Sanctuary Myth

Sanctuary city policies were supposed to create a firewall between local government services and federal immigration enforcement. But data brokers have made that firewall porous. When an undocumented resident signs up for electricity, registers a car, or enrolls a child in school, those records flow to commercial databases that ICE can query without ever contacting the city.[6]

In 2023, the Department of Homeland Security proposed a rule requiring utility companies participating in the Systematic Alien Verification for Entitlements (SAVE) program to share customer data with immigration authorities. The proposal was withdrawn after public outcry — but the commercial data pipelines that accomplish the same thing remain untouched.[7]

Automated Deportation

In January 2025, ICE began piloting an AI system that cross-references social media activity, travel records, and financial transactions to generate “departure probability scores” for individuals on its non-detained docket. The system flags people it predicts are likely to miss court dates — and prioritizes them for arrest.[8]

The algorithm's training data has never been audited publicly. Its error rate is unknown. Its predictions are treated as actionable intelligence. There is no appeals process for a risk score you don't know you have.

The machine doesn't need to see your papers. It already has them.