U.S. Customs and Border Protection (CBP) is advancing its Vehicle Biometric Capability Evaluation (VBCE), pushing its facial recognition infrastructure from pedestrian and air travel environments into the car lanes of land ports of entry.
What began as biometric pilot projects in select pedestrian lanes and airports has now expanded into a sweeping test of facial recognition systems aimed at drivers and passengers in private vehicles.
With the release of CBP’s Privacy Threshold Analysis (PTA) for the VBCE, the contours of this program are clearer. CBP is experimenting with technology that can capture images of every occupant in a moving car and verify their identity against federal databases before they reach an inspection booth.
VBCE is part of CBP’s broader Simplified Arrival program, which is an enhanced international arrival process that uses facial biometrics to automate the manual document checks that are already required for admission into the U.S. It provides travelers a touchless process that further secures and streamlines international arrivals while fulfilling a longstanding congressional mandate to biometrically record the entry and exit of non-U.S. citizens.
For critics, though, the project is another example of how CBP is building an expansive surveillance infrastructure under the guise of efficiency with limited oversight and few guarantees that privacy protections will endure once pilots transition into full-scale operations.
According to the PTA, VBCE relies on a vendor capture system embedded in designated lanes at land ports of entry. As vehicles approach the primary inspection lane, high-resolution cameras capture facial images of occupants through windshields and side windows. The images are then sent to the VBCE platform where they are processed by a “vendor payload service” that prepares the files for CBP’s backend systems.
Each image is stored temporarily in Amazon Web Services’ S3 cloud storage, accompanied by metadata and quality scores. An image-quality service assesses whether the photo is usable while an “occupant count” algorithm tallies the number of people in the vehicle to measure capture rates.
A matching service then calls CBP’s Traveler Verification Service (TVS) – the central biometric database that underpins Simplified Arrival – to retrieve “gallery” images from government holdings such as passports, visas, and other travel documents.
The PTA specifies that an “image purge service” will delete U.S. citizen photos once capture and quality metrics are obtained, and that all images will be purged when the evaluation ends. Still, during the test phase, images can be retained for up to six months, a far longer window than the 12-hour retention policy CBP applies in operational use for U.S. citizens.
The agency emphasizes that participation is voluntary. Lanes designated as “biometric” are marked with signage in English and Spanish, and travelers who decline may choose a standard lane. For those who change their mind mid-queue, additional signage is placed inside the biometric lanes. Local ports are instructed to issue press releases, update wait-time websites and distribute public-affairs guidance so travelers are aware of their options.
CBP has been experimenting with biometric vehicle capture for years, but VBCE represents the most structured attempt yet. The first major test came at the Anzalduas Port of Entry in Texas in 2018, where cameras were installed to capture photos of occupants as cars passed through. A Federal Register notice at the time described it as a proof of concept.
In 2024, CBP conducted a more ambitious test at the Mariposa Port of Entry in Nogales, Arizona where it sought to capture between 85 and 100 percent of passengers in each car before reaching the inspection booth. That test provided technical lessons about lighting, angles, and windshield glare, all factors that complicate automated face capture.
Most recently, during the winter of 2024–2025, CBP deployed the system at the Peace Bridge in Buffalo, New York. For several months, three lanes were marked biometric with cameras mounted to capture every driver and passenger. Travelers who declined could choose a non-biometric lane. CBP said the test would ran into March 2025 to assess performance in cold weather.
These incremental expansions are not simply experiments; they are building blocks. CBP’s July biometrics locations page lists vehicle processing as “in testing,” but at the same time, the agency also issued a Request for Information to industry seeking commercial solutions that can reliably capture facial images in officer-manned primary zones.
The expansion of vehicle biometrics is unfolding against a legislative backdrop that increasingly embeds biometric identity checks into immigration law. The recently passed One Big Beautiful Bill codified an array of surveillance-heavy requirements for immigration enforcement, effectively institutionalizing the use of biometrics across Department of Homeland Security components.
CBP’s Traveler Verification Service has emerged as the operational backbone of this system. TVS now supports facial comparison in airports, seaports, pedestrian crossings, and, with VBCE, it’s now in vehicle lanes. Congress’s move to enshrine these systems in statutes ensures they will outlive pilot programs and become fixtures of border enforcement.
CBP’s biometric expansion is not happening in isolation. In May, DHS announced a series of AI-driven initiatives across Immigration and Customs Enforcement, U.S. Citizenship and Immigration Services, and CBP. Together, they reflect a larger enforcement ecosystem where machine learning underpins every interaction with noncitizens.
VBCE reflects this AI posture. Algorithms score image quality, count occupants, and match faces in real time. CBP officials argue that this increases efficiency and accuracy, reducing officer workload.
But critics see a different pattern. They see that the embedding of AI into frontline immigration enforcement is being done without sufficient checks on bias, error, or mission creep.
While the PTA underscores that participation is voluntary, border dynamics complicate that claim. Opting out often means longer lines, additional questioning, or suspicion of noncompliance. Few travelers are aware they can decline, and signage is easy to miss in the rush of traffic.
Civil-liberties advocates argue that once the infrastructure is in place, the default shifts quickly. Voluntary lanes become mandatory and pilot projects become policy. As the American Civil Liberties Union has warned, CBP’s “face surveillance across airports, seaports, and land borders would put the United States on an extraordinarily dangerous path.”
The Electronic Privacy Information Center has similarly called for a moratorium on biometric deployments until Congress enacts clear rules.
Skepticism about CBP’s assurances is not unfounded. In 2019, a CBP subcontractor improperly copied images collected during a vehicle pilot and later exposed them in a data breach. The DHS Inspector General concluded that CBP had failed to safeguard sensitive data and that its security practices were insufficient to prevent the subcontractor’s actions.
This history looms large over VBCE. While the PTA includes cloud storage protocols and purge mechanisms meant to prevent similar failures, the six-month retention window in testing raises new concerns. For privacy advocates, the breach is evidence that once biometric data is collected, the risk of misuse or exposure never disappears.
Independent watchdogs have repeatedly flagged problems with DHS’ biometric programs. A Government Accountability Office report urged DHS to conduct more rigorous accuracy assessments of its facial recognition systems, particularly for demographic disparities. The Privacy and Civil Liberties Oversight Board has likewise pressed CBP to demonstrate that its systems comply with constitutional protections and privacy law.
Yet, oversight has not kept pace with deployment. GAO noted in 2023 that DHS had not fully implemented its own recommended safeguards, and civil-liberties groups argue that CBP’s transparency measures – press releases, signage, and voluntary lanes – fall short of meaningful accountability.
The VBCE PTA does what it is designed to, which is to identify privacy implications, specify purge rules, and establish communication protocols. But the broader context suggests a familiar pattern. Each pilot has inched CBP closer to making vehicle biometrics routine and each procurement hints at eventual nationwide deployment. Meanwhile, each legislative development lie the One Big Beautiful Bill reinforces biometric enforcement as a central plank of immigration policy.
For CBP, the appeal is obvious. Automated identity verification promises efficiency, accuracy, and enhanced security. But for travelers it signals a shift toward a border experience where every face in every car is scanned, logged, and verified before a window is rolled down.
Civil-liberties advocates warn that without strict limits, the line between border security and domestic surveillance will blur. Whether this trajectory can be reconciled with constitutional protections, privacy rights, and public trust remains unresolved.
Article Topics
biometric identification | biometrics | border security | CBP | data privacy | facial recognition | U.S. Government | Vehicle Biometric Capability
https://www.biometricupdate.com/202509/facial-recognition-moves-into-vehicle-lanes-at-us-borders