ICE’s Facial Recognition App Misidentified a Woman Twice

Our Take

Credit to 404 Media for digging up CBP testimony that confirms what civil liberties advocates have warned for years: facial recognition is not the infallible oracle agencies claim it is. ICE's Mobile Fortify app didn't just stumble once — it misidentified the same woman twice, yet officials have told lawmakers this tool should be trusted over a birth certificate. When an algorithm's guess outranks a government-issued document in the eyes of the agency using it, you have a system built for convenience, not accuracy.

This isn't an isolated glitch, it's a preview of where automated surveillance is headed when it's merged with immigration enforcement. Facial recognition errors disproportionately hit women and people of color, and in this context a bad match doesn't mean a denied loan or a wrongful tag — it can mean detention, deportation, or worse. The same networked, trust-the-algorithm mindset drives the rollout of ALPR systems like Flock Safety's cameras, which are increasingly shared with federal agencies including ICE and CBP, often without the public ever being told their local camera feed is now part of a national dragnet.

If a government app can't reliably tell two different people apart, there's no reason to assume a license plate camera network is any more careful with your data before it's funneled to agencies with far broader mandates than the local police department that installed it. Check our map to see what's watching your neighborhood, and visit our take-action page to push back on these surveillance pipelines before the mistakes become irreversible.