U.S. DHS Plans Unified Biometric System That Merges Face Recognition, Fingerprint Data Across All Enforcement Agencies

Per internal documents reviewed by WIRED, the U.S. Department of Homeland Security (DHS) is moving forward with a plan to consolidate all its existing face recognition and other biometric technologies into a single integrated system. The new platform will be capable of cross-referencing faces, fingerprints, iris scans, and other unique biological identifiers collected across every one of DHS’s enforcement divisions.

DHS has solicited input from private biometric industry contractors to guide development of this unified platform, which will allow authorized staff to search facial and fingerprint records across massive government biometric databases. These databases currently hold biometric data gathered in a wide range of disconnected contexts, and the core goal of the project is to connect systems across all DHS components: Customs and Border Protection (CBP), Immigration and Customs Enforcement (ICE), the Transportation Security Administration (TSA), U.S. Citizenship and Immigration Services, the U.S. Secret Service, and DHS headquarters. The new system will replace the current fragmented patchwork of tools that lack simple, seamless cross-agency data sharing.

Built to support core DHS operations including watch list management, detention, and removal proceedings, the project comes as DHS has rapidly expanded biometric surveillance far beyond official ports of entry, rolling the technology out to intelligence units and undercover agents operating hundreds of miles inland from U.S. borders.

Internal records confirm DHS is seeking to purchase one centralized “matching engine” that can process multiple types of biometric data — faces, fingerprints, iris scans, and more — through a single shared backend, giving all participating DHS agencies access to one common system. In theory, this platform will support both routine identity verification and law enforcement investigative searches.

For face recognition specifically, the system will operate in two distinct modes. Identity verification compares one submitted photo against a single stored identity record, returning a simple yes-or-no match result based on facial similarity. For investigative work, the system searches an entire large database and returns a ranked list of the closest facial matches for a human analyst to review, rather than making a final identification decision on its own.

Both search types carry well-documented real-world technical limitations. Identity verification systems are tuned to be more conservative, so they are far less likely to incorrectly flag an innocent person as a match. However, this design also means they often fail to identify a correct match if the submitted photo is slightly blurry, taken at an odd angle, or outdated. For investigative searches, the matching threshold is set much lower: while this makes it more likely the correct person will appear somewhere in the results, it also produces far more false positive matches that require time-consuming manual review.

Project documents explicitly state DHS wants full control to adjust how strict or permissive the matching threshold is, with customization allowed based on the context and use case of each search.

DHS also requires the new system to integrate directly with its existing agency infrastructure. Contractors will be responsible for connecting the new matching engine to current biometric sensors, identity enrollment systems, and data storage repositories, so information collected by one DHS component can be cross-searched against records held by any other component.

It remains unclear how feasible this full integration actually is. Over decades, different DHS agencies have purchased independent biometric systems from a range of separate private companies. While every system converts a face or fingerprint into a unique string of numerical data for processing, most are engineered to only work with the original vendor’s proprietary software that created them.

In practice, this means a new department-wide search tool cannot simply “flip a switch” to make all legacy systems compatible. DHS will almost certainly need to convert massive volumes of old records into a single standardized format, reprocess records with a new shared algorithm, or build custom software bridges that translate data between incompatible systems. All of these approaches require significant time and funding, and each can impact the final system’s overall speed and accuracy.

At the massive scale DHS is proposing — with a potential total of billions of stored biometric records — even small, overlooked compatibility gaps can quickly escalate into major systemic problems.

Project documents also include an open placeholder noting DHS wants to add voiceprint analysis capability to the system in the future, but no detailed plans are included for how voiceprints would be collected, stored, or searched. DHS has previously used voiceprint technology in its “Alternative to Detention” program, which allowed immigrants to remain in their communities while required to participate in intensive monitoring, including GPS ankle trackers and routine identity check-ins confirmed via biometric voiceprint.

The U.S. Justice Department’s Criminal Resource Manual notes that most federal courts have ruled voiceprint evidence admissible in court, citing cases from the 1970s and 1980s. However, the manual also acknowledges that at least one federal appeals court has found the technique inadmissible, and its scientific validity has remained unconfirmed since the U.S. Supreme Court’s 1993 ruling in Daubert v. Merrell Dow Pharmaceuticals. These questions have only grown more urgent with the rise of AI systems that can convincingly mimic a specific person’s voice.

| Have a Secure Story Tip? |

|---|

| Are you a current or former U.S. government employee with insight into U.S. immigration enforcement? We want to hear from you. Using a personal (non-work) phone or computer, contact the reporter securely via Signal at dell.3030. |

Civil liberties advocates and lawmakers have warned that DHS’s biometric tools are increasingly bleeding into political policing. Already, the agency has photographed and face-scanned U.S. citizens in public spaces during and after political protests, using tools built to identify individuals, map personal connections, and add people to “derogatory” watch lists — all with almost no public transparency and few realistic pathways for people to challenge incorrect or unfair matches.

DHS did not immediately respond to WIRED’s questions about how the proposed unified system will be used for investigative face searches involving U.S. persons, or what safeguards, matching threshold settings, and oversight mechanisms will govern its use across all DHS components.

DHS has never publicly released the privacy rules that govern how its field agents use face recognition, leaving the public unable to access even the most basic guardrails: when agents are allowed to scan a person’s face, what counts as a valid legal reason for a scan, and how long search results are stored in agency systems.

New WIRED reporting published this month found DHS rolled out a mobile face recognition tool called Mobile Fortify in 2023, after rolling back centralized privacy reviews and department-wide restrictions on the technology. At the same time, the agency revoked the original biometric use policy established early in the Biden administration, and has not yet published its own updated replacement, leaving it completely unclear what rules currently govern the technology.

When U.S. Senator Ed Markey introduced the ICE Out of Our Faces Act in early February, he framed face recognition as an enforcement tool no longer confined to official controlled checkpoints. ICE and CBP are using it “to track, target, and surveil individuals across the country,” Markey said, arguing the technology serves not just identification, but intimidation: it can be used to scan people in public spaces to confirm their identity and pull up personal information, even when people are only participating in lawful protest or public criticism of the government.

Markey and the bill’s co-spancers designed the legislation as a direct block on this expanded use: it bars ICE and CBP from acquiring or using face recognition and other biometric identification systems, requires agencies to delete all biometric identifiers they have already collected, and allows individuals and state attorneys general to seek civil penalties for violations of the ban.

Jeff Migliozzi, communications director for the nonprofit Freedom for Immigrants (which operates the National Immigration Detention Hotline), said the expansion of DHS’s biometric infrastructure raises severe, urgent civil rights concerns.

“This development is deeply alarming,” he told WIRED. “These biometric technologies, which have long been riddled with inherent racial biases, are aiding the government’s ability to rapidly expand its surveillance power not only over immigrants and other over-policed communities, but, as we are increasingly seeing with this administration, over political dissidents and anyone in this country regardless of citizenship status.”

“The sheer scale and speed at which big tech is converging its biometric, data, and AI capabilities in direct support of expanding the government’s surveillance and deportation dragnet is incredibly alarming and is a nightmare for civil rights and personal privacy,” Migliozzi added.