ICE's facial recognition app Mobile Fortify incorrectly identified a detained woman twice, returning two different wrong names during an immigration enforcement operation in Oregon, raising concerns about the accuracy of the system used to determine deportation decisions.
In 2024, Immigration and Customs Enforcement (ICE) used its facial recognition app called Mobile Fortify on a 45-year-old woman (MJMA) detained during a raid in Woodburn, Oregon. The CBP official attempted to identify the woman who was refusing to answer questions by taking photos of her face with the app. The first scan returned the incorrect name 'Maria', and when the woman did not respond to that name, officials scanned her face again. The second scan produced a different incorrect name entirely. Both identifications were wrong according to her legal representatives at Innovation Law Lab. The woman was later released from an ICE detention center in Tacoma without conditions the next day. She had entered the U.S. with a B-2 tourism visa and is seeking asylum. ICE has previously told lawmakers that Mobile Fortify provides a 'definitive' determination of immigration status and should be trusted over birth certificates. The app uses CBP systems designed for border entry verification but has been deployed domestically on American streets, accessing databases of over 200 million images. The incident highlights accuracy concerns with facial recognition technology, which research shows performs particularly poorly on women of color.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI systems that fail to perform reliably or effectively under varying conditions, exposing them to errors and failures that can have significant consequences, especially in critical applications or areas that require moral reasoning.
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed