AI now screens travelers and cargo at the border. AVAAS certifies the system where the screening decision reaches a person.
DHS, CBP, and TSA screen travelers, cargo, and passengers through automated systems and match faces against federal databases. A wrong flag pulls a person aside or holds a shipment. AVAAS certifies the system at the point its output drives a screening or enforcement decision.
The decision point at the border
In screening, the AI acts when it scores a traveler or shipment for risk, matches a face, or flags an anomaly. A wrong output sends a real person to secondary inspection or holds legitimate cargo.
DHS's own directive prohibits AI as the sole basis for an enforcement action. The system still has to be tested to show it holds to that.
What keeps components exposed
High-impact AI on a clock
OMB M-25-21 names screening and biometric identification high-impact, and requires pre-deployment testing, monitoring, and documentation, or the system has to be discontinued. M-25-22 extends pre-award testing and validation to the AI agencies buy.
A human decision-maker is required
DHS Directive 139-08 bars AI as the sole basis for an enforcement action and requires a human decision-maker. Demonstrating that the system actually holds to that takes independent evidence, not assurances.
Identity matching across the population
ICE's Mobile Fortify draws on more than 200 million images from DHS, FBI, and State Department databases. Face recognition misidentifies people of color at higher rates, and DHS removed its face-recognition directive from public posting in early 2025.
Evidence the system meets its federal duties
Does it meet high-impact testing duties?
AVAAS produces documented, third-party evidence aligned to the pre-deployment testing M-25-21 and M-25-22 require for high-impact AI.
Is a human actually in the loop?
AVAAS evaluates whether human oversight is meaningful at the point the output drives an action, not just present on paper.
Does it misidentify across groups?
Five structurally independent validators test for demographic disparity and failure patterns using causal attribution.
You get documented, third-party evidence that a screening or identification system meets the testing federal high-impact AI rules now require.
Related AVAAS coverage: Certification · Evidence Ledger. Or run the free AI Exposure Assessment to see what applies to you.
See where your screening AI creates liability.
Tell us where automated screening or identification informs a decision about a person, and we will scope an AVAAS certification to the exposure.
Ready to start now? Certify Your AI → or email [email protected]