ICE using data and probability to decide where to detain and arrest people

U.S. Immigration and Customs Enforcement’s Enhanced Leads Identification & Targeting for Enforcement (ELITE) tool is being used to identify “targets” and to direct enforcement activity as part of a larger, heavily funded analytics ecosystem built by Palantir Technologies. The problem is that relying on probabilistic “confidence scores” raises fundamental legal questions about warrants, probable cause, and the limits of lawful arrest authority.
This system favors neighborhood-scale deployment over address-specific certainty, creating a persistent tension between operational efficiency and constitutional safeguards.
ELITE is embedded within ICE’s broader data-fusion, analytics, and case-management infrastructure built by Palantir, converting administrative records into actionable leads for directed enforcement actions.
It functions as a geospatial interface rather than a simple lookup tool, and allows officers to visualize, prioritize, and select targets based on location, identity data, and assessments of where individuals are likely to be found.
By translating fused backend records into map-based deployment strategies, the system treats identity data, address confidence, and location density as core operational variables to guide immigration arrests by presenting potential enforcement subjects as points on a digital map.
Selecting an individual opens a dossier on that person which contains biographic identifiers such as name, date of birth, Alien Registration Number, and photograph.
The system assigns a numerical confidence score, expressed on a scale of zero to 100, indicating how likely it is that the individual is currently located at the associated address.
That score is used operationally to determine whether an address is worth attempting to serve an administrative arrest warrant and, more broadly, to identify geographic areas where multiple viable targets appear to be clustered.
A core feature of ELITE is its geospatial lead-sourcing capability. Officers can select targets individually or draw a boundary around a neighborhood-scale area and return multiple potential subjects at once.
Results can be filtered using categories that include biographic identifiers, criminal history indicators, location attributes, and operational parameters. This design allows the system to be used not only to locate specific individuals, but to identify what internal ICE documentation has described as “target-rich” areas where enforcement resources can be concentrated for maximum yield.
Address and identity information surfaced through ELITE is drawn from multiple federal data sources.
The reliance on non-law enforcement administrative data reflects a broader evolution in immigration enforcement targeting, as ICE increasingly depends on inter-agency data sharing and centralized analytics rather than direct cooperation from state or local law enforcement agencies.
The use of probabilistic confidence scoring has important legal and operational implications. While the Fourth Amendment does not require absolute certainty to support a judicial search warrant, it does require individualized probable cause grounded in specific, articulable facts.
A numerical likelihood score generated by a data system does not, on its own, establish that a person is present inside a particular residence at a given time. As a result, ELITE’s confidence scores are not well suited to satisfying the constitutional standards required for obtaining judicial warrants to enter homes.
In practice, this limitation helps explain why ELITE appears designed to support area-based enforcement rather than address-specific home entries. ICE agents generally do not rely on judicial search warrants to carry out civil immigration arrests.
Instead, they operate under administrative arrest warrants and conduct enforcement actions in public or quasi-public spaces, such as sidewalks, parking lots, apartment courtyards, and building thresholds, where constitutional protections are more limited.
By identifying neighborhoods where targets are statistically more likely to be encountered, ELITE enables agents to position themselves in those areas and wait for individuals to emerge, rather than attempting to prove that a specific person is inside a specific residence.
From this perspective, ELITE’s confidence scoring is less about establishing certainty than it is about guiding deployment. The system allows ICE to decide where to apply enforcement pressure without needing to test the reliability of its data before a judge.
The practical consequences of that enforcement model have become increasingly visible on the ground. In Minneapolis this week, amid escalating protests following the shooting of a man during an encounter with federal agents, Department of Homeland Security Secretary Kristi Noem publicly defended ICE officers who were demanding that people nearby prove they were U.S. citizens.
Speaking to reporters outside the White House, Noem framed the encounters as a routine extension of “targeted enforcement,” even when they involved individuals not named in any arrest order.
“In every situation we are doing targeted enforcement,” Noem said. “If we are on a target, and doing an operation, there may be individuals surrounding that criminal that we may be asking who they are and why they’re there and having them validate their identity.” She added that officers may detain and question those individuals while “we run that processing.”
The remarks came amid mounting reports of federal agents detaining and questioning U.S. citizens not only in Minnesota but in other states as well, raising fresh scrutiny of the legal thresholds governing immigration stops. Former federal prosecutor Elie Honig said Noem’s description overstated the authority immigration officers possess in public spaces.
“It’s wrong. It’s illegal, and it’s unconstitutional to require people to show their citizenship papers without some other basis to make a stop,” Honig said.
While immigration officers operate under a lower standard than criminal investigators, he noted that the Constitution still requires reasonable suspicion – articulable facts that can be explained to a court – to justify even brief detention or questioning.
Honig pointed to a recent Supreme Court decision in which Justice Brett Kavanaugh, writing a concurring opinion in September 2025, reaffirmed that immigration officers must meet that standard before stopping or detaining someone, even temporarily.
“You cannot just go arbitrarily up to people or set up a checkpoint or go door to door and say, ‘Hey, you need to prove to us that you’re a U.S. citizen,’” Honig said.
That legal boundary becomes especially strained in area-based enforcement operations. If officers rely on neighborhood-level targeting rather than address-specific certainty, encounters are more likely to involve bystanders whose only connection to an operation is proximity.
Honig said it would be difficult to justify such stops under existing precedent.
“Why would standing around an ICE operation tend to indicate that somebody is in this country illegally?” he asked. “That is quite a stretch.”
The tension underscores a central question raised by ICE’s modern targeting systems: when enforcement decisions are guided by probabilistic assessments of where people might be, rather than concrete evidence of who is present at a specific location, the line between lawful inquiry and unconstitutional detention becomes harder to define, and easier to cross.
From a procurement standpoint, ELITE appears to be embedded within ICE’s larger Investigative Case Management (ICM) ecosystem. That ecosystem is supported by a long-running Palantir Technologies delivery order issued under GSA Schedule GS35F0086U, delivery order number 70CTD022FR0000170, which covers ICM operations and maintenance support as well as custom system enhancements.
Public federal procurement records show that this delivery order has expanded steadily through a series of large modifications.
Obligations recorded in the Federal Procurement Data System include $1.2 million in January 2023; $19.3 million in September 2023; $1.7 million in August 2024 and $18.6 million in September 2024.
There was a nominal $2,499 action in March 2025; a nearly $30 million modification in April 2025; a negative adjustment of $18,000 in July 2025; $19.3 million in September 2025; another $29.8 million modification later that same month, and an additional $1,961,877 action in late September 2025.
Contract tracking platforms list the total potential value of the delivery order at nearly $160 million, with total obligations approaching $140 million.
Within that same contract stream, ICE has funded the development and sustainment of the Immigration Lifecycle Operating System (ImmigrationOS), which is an operating system-level environment layered on top of ICM intended to support lifecycle tracking, enforcement prioritization, and advanced analytics.
The recurring $29.8 million modifications associated with the delivery order have been tied to continued licensing, operations and maintenance, and adaptive maintenance for ImmigrationOS.
ELITE fits within this environment as a tactical targeting layer. While ImmigrationOS and ICM provide case management, data integration, and prioritization across the immigration lifecycle, ELITE translates those underlying records into a field-usable interface that supports real-time deployment decisions. It operationalizes ICE’s analytic back end.
That back end includes the Data Analysis System (DAS), an ICE Enforcement and Removal Operations database designed to aggregate and analyze information to help locate individuals subject to removal. DAS functions as an analytical reservoir, fusing data from multiple sources and making it available for downstream use.
ELITE appears to consume outputs from such systems, presenting them visually and pairing them with confidence metrics that guide enforcement action.
ELITE also aligns with the mission of ICE’s Targeting Operations Division, which includes the Law Enforcement Support Center, the National Criminal Analysis and Targeting Center, and the Pacific Enforcement Response Center.
These centers generate intelligence-driven leads for field offices by applying analytical and technical capabilities at scale. ELITE serves as a bridge between centralized lead generation and street-level operations, enabling officers to convert abstract leads into geographically bounded enforcement plans.
Taken together, ICE has and continues to build out a mature enforcement architecture built around continuous data ingestion, analytics, and geospatial decision-making. But the system’s emphasis on probabilistic location assessment and area-based deployment underscores a broader shift in immigration enforcement which increasingly appears to be on shaky legal ground.
Article Topics
DHS | ELITE | ICE | law enforcement | location data | Palantir | surveillance | U.S. Government







Comments