ICE to Pay LexisNexis Millions for Data to Feed Palantir's System

U.S. Immigration and Customs Enforcement (ICE) plans to pay LexisNexis millions of dollars to keep accessing its data, which the agency intends to load into the Palantir platform. That's according to recently published procurement documents reported on by the outlet 404 Media. ICE estimates the contract will cost roughly $6.7 million.

LexisNexis has long supplied data to law enforcement across different administrations. At issue is access to the LexID and Accurint Virtual Crime Center tools, which — by the company's own account — tie together more than 82 billion public and proprietary records from over 10,000 sources, helping police locate specific individuals.

Data for deportations

According to the documents, the information is meant to support “all aspects of vetting and screening, lead development, and criminal analysis” at ICE. Among other things, the data is intended for the Enforcement and Removal Operations (ERO) division, which handles deportations. The news comes as ICE detained more than 51,000 people in July alone, many of them at airports.

The document lays out an explicit requirement: the database must interface via API with ICE applications, “including but not limited to the Palantir platform, PenLink, and ICE Data Analytics.” PenLink is another ICE contractor that sells social media monitoring tools and Webloc, a product for tracking the location of mobile phones across entire neighborhoods.

Palantir has become a key provider of technical infrastructure for the mass deportation program. Earlier, 404 Media revealed the existence of ELITE (Enhanced Leads Identification & Targeting for Enforcement), a system Palantir built for ICE: it maps potential targets and gives officers a “confidence score” for a person's current address, drawing in part on data from the Department of Health and Human Services. Whether the LexisNexis data will be loaded specifically into ELITE or into another Palantir platform is unclear from the documents.

AI and facial recognition requirements

The document separately lists requirements tied to artificial intelligence. ICE wants an “AI-powered identification system” capable of “inferring the identity of the end user” from the data, as well as a facial recognition system that uses “large-scale image databases for high-accuracy matching of faces from diverse sources, including open-source material.” All of it is supposed to run in bulk, streaming mode.

LexisNexis, Palantir, and ICE did not respond to the outlet's requests for comment.

Source: 404 Media