The artificial intelligence (AI) tools built to guard America’s borders are now extending their policing powers into America’s neighborhoods. What began as AI-powered immigration enforcement confined primarily to our southern border is rapidly becoming something else: a quiet expansion of government surveillance into everyday life.
Facial recognition, biometric scanning, and social-media monitoring — once justified for tracking noncitizens — are now being used to identify and investigate U.S. citizens. This is mission creep, a military term for a shift in objectives that gradually turns a limited mission into something far larger. And it’s happening right now.
The Department of Homeland Security (DHS) is accelerating its investment in artificial intelligence for U.S. Immigration and Customs Enforcement (ICE) with little public scrutiny. Procurement records reveal a sweeping expansion of surveillance technology, including:
$30 million for Palantir’s ImmigrationOS, which promises “granular tracking” of immigrants, including real-time monitoring of self-deportations.
$4.6 million for iris-scanning smartphones from BI2 Technologies.
$3.75 million for a Clearview AI facial-recognition contract — ICE’s largest purchase of the technology to date.
This is not just abstract policy. These surveillance technologies are already being used on the streets. Internal footage obtained by Media 404 shows ICE officers using facial recognition apps to determine the citizenship status of teenagers who were not carrying IDs. One app, Mobile Fortify, draws from over 200 million images stored in the databases of DHS, the FBI, and the State Department. It also appears to tap into super query tools, systems that aggregate data across multiple government databases for deeper access into personal networks.
Since Media 404’s initial reporting, additional investigations have made clear that these were not isolated uses. Reporting by Minnesota Public Radio and the Guardian shows ICE acquiring and operating AI-enabled surveillance tools that surveil U.S. citizens and expand beyond the reach of immigration enforcement. In Minnesota, ICE has purchased and deployed social media monitoring and location-tracking systems that allow agents to analyze the movements of large groups of people in specific locations — capabilities that, according to company materials and reporting, are explicitly marketed for use at protests and other gatherings protected under the First Amendment. ICE and other agencies haven’t specifically said who they are targeting, but these tools let them monitor where people have been and map their connections — without a warrant or specific suspicion. Increased AI surveillance comes at a time when ICE agents have already been instructed to collect the personal information of protestors, as they did with Alex Pretti days before his murder in Minneapolis.
These represent new overreaches of surveillance. What began as tools for immigration enforcement are now positioned for domestic political surveillance, sweeping Americans into massive data monitoring systems that they were never told even existed.
This is mission creep, and it is dangerous. Tools built and justified to efficiently tackle one issue quietly scale up, in this case collapsing the boundary between targeted enforcement and population-wide surveillance. Here mission creep is reshaping domestic policing. ICE Homeland Security Investigations (HSI), a division originally focused on cross-border threats, is now invoking domestic terrorism concerns to monitor activity within the United States, including public protests. Paired with powerful AI surveillance technology, ICE and other agencies have now essentially been given broad latitude to spy on all Americans.
Critics are sounding the alarm. Sen. Ron Wyden (D-OR) warned that ICE could “trample on the rights of Americans and anyone Trump labels as an enemy.” Maria Villegas Bravo of the Electronic Privacy Information Center (EPIC) flagged likely violations of the First and Fourth amendments. Oversight has eroded as well. Internal watchdogs have been sidelined or shut down, and DHS continues to operate without a comprehensive biometric privacy law or a mandatory, agency-wide AI audit requirement. In February, Sen. Ed Markey (D-MA) introduced legislation to ban DHS and ICE from using facial recognition and other biometric surveillance inside the United States, underscoring growing concern in Congress that these tools are being deployed far beyond their original immigration mandate.
Homeland security says the expanded surveillance measures are justified to boost efficiency and safeguard national security. But that language masks a deeper reality: the systems designed to track immigrants can easily be flipped to scan entire populations. Tools built for speed, scale, and efficiency make population-wide surveillance nearly effortless. Unified databases and automated analytics quickly turn efficiency into overreach. Few companies illustrate this transformation more clearly than Palantir Technologies, a data analysis software company with billions of dollars in government contracts.
Palantir’s ImmigrationOS doesn’t just track immigration cases. It manages the immigration lifecycle process of individuals through predictive analysis, surveillance, and enforcement. The same language that describes efficient deportations could just as easily be used to justify mass surveillance of any flagged group.
Before this mission creep becomes accepted policy, we need to establish clear limits. This means clarifying when a warrant is required for law enforcement to conduct biometric scans of citizens and noncitizens, a question that remains inconsistently addressed across jurisdictions. Furthermore, to ensure accountability, AI surveillance tools deployed in the field should be subject to public reporting requirements, and independent auditors must regularly review platforms like ImmigrationOS and Mobile Fortify to address public concerns. At the same time, Congress should advance comprehensive data-privacy protections and civil rights legislation that keep pace with the rapid expansion of AI.
Without these safeguards, immigration enforcement risks becoming a testing ground for broader domestic surveillance — one where powerful technology normalizes suspicion, automates targeting, and erodes constitutional protections. The line between border enforcement and everyday policing is already fading. Whether it disappears entirely depends on policymakers, who must act now to reassert transparency, accountability, and democratic oversight.