ICE's Face App Gave Her Two Wrong Names

The agent said the match was "a maybe." He detained her anyway. DHS had already removed its own safeguards.

TL;DR: Court testimony from Oregon reveals that ICE's Mobile Fortify app returned two different wrong identities when agents scanned the same woman's face twice. The agent admitted the match was rated "a maybe." His probable cause? She spoke Spanish. DHS quietly scrubbed its own facial recognition safeguards from its website three weeks after Trump's inauguration. The app has been used over 100,000 times. It generates guesses, not confirmations, but agents are treating those guesses as grounds for detention.

Two Scans. Two Wrong Answers. One Detention.

During an immigration operation in Oregon, a CBP agent scanned a handcuffed woman's face with Mobile Fortify. The app returned a name. The agent scanned her again. A different name came back.[1]

Both names were wrong.

The woman, identified in court filings as "MJMA," a farmworker with a valid B-2 visa who was seeking asylum, yelped in pain as the agent physically repositioned her head to get a better angle for the camera. The first result came back rated "a maybe." The second: "possible."[1][2]

When asked in court about the confidence level, the agent testified: "I don't know."

His own assessment of the app's output? "It's just an image, your honor. You have to look at the eyes and the nose and the mouth and the lips."[1]

Despite having two contradictory results from the same person, the agent cited three reasons for probable cause: the woman spoke Spanish, she was near people who appeared to be noncitizens, and Mobile Fortify returned a "possible match."[1]

Fingerprints (a far more reliable biometric) were never taken. They could have been. ICE collects them in the office all the time. But the agent didn't bother.[1]

MJMA was detained overnight and released the next day without conditions.

The App Generates Guesses. ICE Treats Them as Facts.

Here's what every facial recognition vendor, every police department with a policy, and every independent researcher agrees on: facial recognition cannot positively identify anyone.

Nathan Wessler, deputy director at the ACLU, put it plainly: "Every manufacturer of this technology, every police department with a policy makes very clear that face recognition technology is not capable of providing a positive identification, that it makes mistakes, and that it's only for generating leads."[1]

Mobile Fortify generates candidate matches, not confirmations. It compares a photo taken in the field (bad lighting, wrong angles, motion blur, handcuffed subjects looking at the ground) against a database of 1.2 billion government images taken under controlled conditions (passport photos, visa applications, mugshots).[3]

The gap between controlled studio photos and a picture snapped on a phone in a parking lot is enormous. And that gap is where people get detained.

Mario Trujillo, senior staff attorney at the Electronic Frontier Foundation, flagged the missing safeguards: "Facial recognition can be wrong, and it has been wrong in the past. Here, the safeguards you'd expect (confidence scores, clear thresholds, multiple candidate photos) don't appear to be there."[1]

DHS Removed Its Own Rules. Three Weeks After Inauguration.

In 2023, DHS published Directive 026-11, its internal policy governing facial recognition use. The directive explicitly prohibited:[1]

  • Using facial recognition as the sole basis for law enforcement or civil enforcement actions
  • Systemic, indiscriminate, or wide-scale monitoring of people
  • Tools that claim to analyze faces for personal traits
  • Deploying facial recognition without headquarters-level review by DHS's chief privacy and information officers

Three weeks after Trump's inauguration, the directive vanished from the DHS website.[1]

Jeramie Scott, from the Electronic Privacy Information Center, confirmed: "DHS has not publicly replaced its previous facial recognition directive, and it appears DHS has no policy or even restrictions on the use of facial-recognition technology."[1]

The person now in charge of reviewing facial recognition privacy? Roman Jankowski (a former Heritage Foundation lawyer and contributor to Project 2025) was appointed DHS Chief Privacy Officer on inauguration day. By March 3, 2025, CBP had "assumed responsibility for the review and adjudication" of privacy reviews under his office's guidance.[1]

The fox is guarding the henhouse. And the henhouse doesn't have a lock anymore.

100,000 Uses. Zero Public Accountability.

According to a federal lawsuit filed by the State of Illinois and City of Chicago, Mobile Fortify has been used "in the field over 100,000 times" since its launch in spring 2025.[1][4]

100,000 face scans. How many were wrong? Nobody knows. ICE doesn't track error rates. There's no public accuracy report. No audit. No oversight mechanism.[1]

The data from each scan doesn't disappear after the encounter. Photos get stored across multiple systems:[1]

ATS

Automated Targeting System: stores photos up to 15 years, regardless of whether you're a citizen or were correctly identified

TVS

Traveler Verification System: cross-references entry/exit records with facial images from airports and ports

IDENT

DHS's massive biometric database: fingerprints stored for a minimum of 75 years. 270+ million records and growing

SAW

Seizure and Apprehension Workflow: records every encounter, whether it leads to arrest or not

Even if Mobile Fortify gets your identity wrong, your face stays in the system for up to 15 years. Wrongly scanned? Too bad. Your biometric data feeds the 1.2-billion-image database that will be used to scan the next person.[3]

Local Cops Know Better. ICE Doesn't Care.

The irony is painful. Police departments across the country already know facial recognition can't verify identity. Their own policies say so:

"A facial recognition match is merely a lead; it is not probable cause."
NYPD Commissioner Dermot Shea[5]
"We don't make an arrest because an algorithm tells us to."
Miami Assistant Chief Armando Aguilar[5]

Philadelphia PD's policy states that matches are "NOT indicative of a positive identification." Los Angeles, Detroit, and Orlando have similar rules. More than half a dozen states have legally codified the principle that facial recognition matches aren't proof of anything.[5]

But ICE? Rep. Bennie Thompson (D-MS) warned that "an ICE officer may ignore evidence of American citizenship (including a birth certificate) if the app says the person is an alien."[6]

An algorithm's guess outranks a government-issued document. That's where we are.

What This Means for You

More than 170 US citizens have been detained by immigration agents since January 2025.[6] The app doesn't distinguish between citizens and noncitizens: it scans everyone and returns guesses for everyone.

NIST studies confirm that facial recognition performs worse on field photos than controlled conditions, and demonstrates reduced accuracy on Black individuals and women.[5] Mobile Fortify is being deployed disproportionately in communities where these error rates are highest.

Congress has introduced two bills to address this: Rep. Thompson's Realigning Mobile Phone Biometrics for American Privacy Protection Act, and the ICE Out of My Face Act from Senators Markey, Merkley, and Wyden.[6][7] Neither has passed. Neither has Republican support.

What You Can Do

Know the Score

Mobile Fortify doesn't verify identity. It generates guesses. If an agent tells you "the app confirmed your identity," that's not how the technology works. A match is a lead, not proof.

Carry Documentation

A passport or birth certificate creates a record, even if agents claim the app overrides documents. Having proof of citizenship matters in court afterward.

Record Encounters

If safe to do so, record any interaction with ICE agents using Mobile Fortify. Note badge numbers, time, location, and what was said. Contact the ACLU or local legal aid.

Contact Your Representatives

Ask your members of Congress to co-sponsor the ICE Out of My Face Act and the Realigning Mobile Phone Biometrics Act. Both need bipartisan support to pass.

References

  1. New York Times: ICE and CBP's Face-Recognition App Can't Actually Verify Who People Are (February 5, 2026)
  2. 404 Media: ICE's Facial Recognition App Misidentified a Woman. Twice. (February 2026)
  3. Bloomberg: DHS Face-Scanning App Pulls From 1.2 Billion-Image Database (February 2, 2026)
  4. PBS: DHS Intensifies Surveillance in Immigration Raids, Sweeping in Citizens (February 2026)
  5. TechPolicy.Press: ICE's Reckless Reliance on Facial Recognition Puts Us All In Danger (February 2026)
  6. ACLU: Face Recognition and the 'Trump Terror': A Marriage Made in Hell (February 2026)
  7. Rep. Jayapal: ICE Out of My Face Act (February 5, 2026)