Palantir ImmigrationOS — ICE Pays $30M for AI System Creating Neighborhood Deportation Maps (2026)

Incident Summary

US Immigration and Customs Enforcement (ICE) contracted Palantir for $30 million to deploy ImmigrationOS, an AI platform that creates neighborhood-level maps for deportation targeting operations.^[1] The system aggregates data from multiple sources to generate geographic targeting maps that identify neighborhoods with high concentrations of potential deportation targets, enabling ICE to plan enforcement operations with AI-driven precision.^[2] Separately, an ICE AI recruitment tool was found to flag anyone with the word “officer” on their resume as having law enforcement experience — a basic keyword-matching error that demonstrates the low quality of some AI tools in the immigration enforcement ecosystem.^[1] Civil liberties organizations including the ACLU and American Immigration Council have raised concerns about minimal transparency regarding the system’s bias, accuracy, and due process protections, noting that individuals targeted by AI-generated deportation maps have no ability to know they have been flagged or to contest their inclusion in targeting lists.^[3]

Key Facts

• Contract: $30 million ICE-Palantir contract for ImmigrationOS^[1]
• Function: Creates neighborhood maps for deportation targeting^[2]
• AI errors: Recruitment tool flags “officer” on resume as law enforcement experience^[1]
• Transparency: Minimal disclosure on bias, accuracy, or due process^[3]
• Concerns raised by: ACLU, American Immigration Council^[3]

Threat Patterns Involved

Primary: Mass Surveillance Amplification — ImmigrationOS amplifies ICE’s surveillance and enforcement capabilities by transforming aggregate data into neighborhood-level deportation maps, enabling a form of geographic targeting that would be infeasible without AI-driven data aggregation and pattern recognition.

Significance

1. AI-powered geographic targeting — The creation of neighborhood deportation maps represents AI-enabled geographic targeting of communities, raising concerns about whether entire neighborhoods are subjected to heightened enforcement based on aggregate demographic data rather than individualized suspicion
2. $30M investment signals scale — The $30 million contract value indicates that AI-powered immigration enforcement is being deployed at significant scale, with the investment suggesting long-term infrastructure rather than a pilot program
3. Due process vacuum — The absence of transparency regarding how individuals are flagged by ImmigrationOS means that affected people have no ability to know they are targeted, no mechanism to contest their inclusion, and no visibility into the data that generated their profile
4. AI quality concerns — The recruitment tool’s “officer” keyword error demonstrates that AI tools in the immigration enforcement ecosystem operate at varying quality levels, raising questions about whether the more consequential targeting algorithms in ImmigrationOS contain similar fundamental errors