AI summary
Summary of article:
Experts say ICE is using AI to combine government and private data to track immigrants. Everything from welfare, housing to social media can become clues and lead to privacy, legal and mistaken identity disputes.
New York University Recently held “Artificial Intelligence andmigrant Many experts on the “AI and Immigration Policy” forum pointed out that as the federal government is accelerating the introduction of artificial intelligence (AI) into immigration enforcement, Immigration and Customs Enforcement (Immigration and Customs Enforcement).snow ) is using AI to integrate massive amounts of data from various government agencies and private companies, including Medicaid, SNAP, public housing and voter registration records, to help identify, locate and track immigrants.
The forum is hosted by the NYU Migration Network. Steven Hubbard, a senior data scientist at the American Immigration Council, said AI is rapidly changing the way federal immigration enforcement is conducted. In the past, data search and comparison work that required a lot of manpower and took several years can now be completed in a short time, greatly improving the efficiency and scale of implementation.
“Using AI to integrate data from different organizations”
From July 2025 to January 2026, there were 58 new immigration-related AI use cases, according to the agency’s analysis of public information from the Department of Homeland Security (DHS), of which ICE added 28 and Customs and Border Protection (CBP) added 23. The two law enforcement agencies accounted for nearly 90% of the new cases. Half of the new systems use generative AI, which can be used to compile investigation data, search records and analyze cases.
Hubbard pointed out that what is noteworthy is not only the use of more AI by the government, but also that the technology mainly flows to “immigration enforcement” rather than public services such as visa and immigration applications. He said the Department of Homeland Security already has entry-exit, travel and biometric records, and now gets the information through other government agencies and commercial data providers like LexisNexis. AI makes it easy to link records from different sources to identify individuals and their relationships.
Saira Hussain, senior staff counsel at the Electronic Frontier Foundation (EFF), said the federal government is seeking to consolidate information held by the Internal Revenue Service (IRS), the Department of Health and Human Services (HHS) and state governments, including white card, food stamp, public housing and voter registration records. However, this does not mean that all welfare recipients’ information has been used by ICE for law enforcement.
He specifically pointed out that some of the information sold by private data brokers is so detailed that if the government collected it itself, it might need to obtain a search warrant. But now the government can obtain it through commercial purchases, which would create a legal loophole in the Fourth Amendment’s protection of individual privacy.
“Google Maps for Immigration”
Hussain also introduced the ELITE system developed by technology company Palantir. The tool displays potential immigration enforcement targets through maps and aggregates relevant personal data to help ICE determine where to prioritize operations. He quoted ICE personnel in the lawsuit as describing the technology as “like the immigration version of Google Maps.”
Katya Schwenk, a reporter for investigative news media The Lever, pointed out that ICE is increasingly relying on private technology companies to provide surveillance and AI tools. For example, they awarded a contract worth approximately NT$5.7 million to Zignal Labs to provide AI social media monitoring services; Another company, Edge Ops, received a non-competitive contract worth approximately NT$12 million to develop an immigration tracking system.
Schwenk said his investigation found that some of the photos of senior officers on the Edge Ops website came from a gallery, and the identities of the personnel in question could not be verified. He reminded that the government is eager to purchase new technologies, but it does not fully understand whether the AI capabilities claimed by suppliers are true.
Apart from tracking location, AI is also used to monitor online speech. Hussein said the EFF is participating in a lawsuit filed by three unions that accuses the federal government of monitoring the social media activities of non-citizens to identify political views deemed “unpopular” by the government, including those supporting Palestine or critical of U.S. policies. The court has rejected the government’s motion to dismiss the case, allowing the trial to proceed.
Hubbard concluded by pointing out that the government often emphasizes that AI is only an assistive tool and that ultimately humans still make decisions, but the real question is whether law enforcement officials understand the data used by AI and its potential errors. He believes that when AI generates inaccurate information and affects the fate of immigrants, whether relevant parties can raise questions and request corrections is the main issue this system must face.
Summary FAQ
-
The text says ICE is integrating white cards, food stamps, public accommodations, voter registration, entry and exit records, and business information, and using AI to compare different sources to establish individuals and their kinship networks.
-
Because the government can obtain detailed information through data brokers that might otherwise require a search warrant, this creates a Fourth Amendment controversy; If the AI makes wrong predictions or the data is inaccurate, it may be difficult for the parties involved to quickly challenge and correct it.
-
Incorporating Palantir’s elite mapping system, Zignal Labs’ social media monitoring service, and Edge Ops’ immigration tracking system, they show that ICE is increasingly relying on private technology companies to provide surveillance and analysis tools.
