Traveler Verification Service met its accuracy goals at airport exits

GAO reports that a 2019 CBP test found that face matching at departure gates correctly matched 98% of travelers' photos, while airlines photographed only about 80% of travelers.

  • 2 an organization

    The test measured 98% correct matches and 0.0092% false matches, which lets CBP record departures as biometrically confirmed.

Depth runs from 1, a convenience or time saved, to 5, a life or a livelihood preserved.

Status
Reported and intended benefits.
Evidence

foundstudiedreportedclaimed

An independent audit report supports the benefit.

Compared with

The minimum accuracy requirements in CBP's Operational Requirements Document. Measured.

Without the system
At least 90 percent correct matches and at most 0.1 percent false matches
With it
98 percent correct matches and 0.0092 percent false matches
The system’s part
Not isolated. The accuracy figures come from CBP's own test agent, and GAO checked the calculations. GAO notes that NIST was to test an algorithm similar to the one in the service for demographic effects. GAO does not name the algorithm's supplier or say how it was trained.
Limits
Not yet checked.
Costs and harms
Not yet checked.
Later
Expanded. GAO reports that DHS leadership approved air exit in December 2019 to move forward and fully deploy to airports.

Download data Cite Suggest a correction

Benefit type
Improved Task Accuracy
System
Traveler Verification Service (TVS), facial recognition for biometric air exit
Made by
U.S. Customs and Border Protection (GAO does not name the algorithm's supplier)
Used by
U.S. Customs and Border Protection with airline and airport partners
Where
Detroit, New York JFK and Orlando airports, United States (test sites)
Stage
In a pilot
Date
Kind
Aggregate result
Broad benefit
Identifying Objects, Actions, and Events
Collections
Scrutiny and findings
Last revised
31 October 2026

Sources

  1. Audit report, by authors with no tie to the system.

    U.S. Government Accountability Office. Facial Recognition: CBP and TSA are Taking Steps to Implement Programs, but CBP Should Address Privacy and System Performance Issues. GAO-20-568, 2 September 2020, pp. 50-52.

    Operational testing showed that air exit was able to correctly match 98 percent of travelers’ photos with photo galleries built from passenger manifests, a key capability for the program.

Details

AI system

Workflow

Galleries for each flight were built from passenger manifests and photos in government records. Airline staff photographed travelers at the gate with cameras linked to the service.

Follow-up

Conditions for replication

Cameras and gate systems must work and be set up correctly. Airlines must photograph all in-scope travelers and avoid manual boarding.

Entry details

Sector

Border control (exit records).

Purpose

U.S. Customs and Border Protection built the Traveler Verification Service to match a traveler's live photo at the gate with photos of the passengers on the flight.

Not yet checked: model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, AI contribution, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.

Cite this record

Peter H. Vartanian. (2026). Traveler Verification Service met its accuracy goals at airport exits (AIBD-0676). Artificial Intelligence Benefits Database, edition of 31 October 2026. https://petervartanian.xyz/aibd/records/AIBD-0676.html (archived at https://doi.org/10.7910/DVN/AADSII, version 4.0)

BibTeX Data Field definitions