US gives local police a face-scanning app similar to one used by ICE agents (arstechnica.com)

US Customs and Border Protection quietly released an Android app called Mobile Identify on Oct. 30 that lets local law-enforcement agencies supporting federal immigration operations use face-scanning tools in the field. The Google Play listing says the app implements functions authorized by Section 287(g) of the Immigration and Nationality Act — meaning participating agencies that sign a DHS Memorandum of Agreement and certify officers can run immigration-ID workflows. A screenshot shows the app requests camera access “to take photos of subjects.” Reporting by 404 Media, which downloaded and decompiled the APK, found code references to “facescanner,” “FacePresence” and “No facial image found.” According to the outlet, the app does not directly return names after a face search; it either provides a reference number for ICE to follow up or advises officers not to detain the person based on the result. The app is currently Android-only. For the AI/ML community, this rollout matters because it extends biometric face-scanning into more distributed, locally run workflows, increasing scale and operational risk even if identifying decisions are deferred to ICE. The decompiled artifacts suggest on-device camera capture and face-detection logic, but it’s unclear if matching happens locally or via a remote database — a key technical distinction for latency, privacy and threat models. Broader implications include governance and auditability of deployed ML models, bias and false-positive risks in field-facing systems, and legal/ethical concerns around delegating biometric surveillance to municipal agencies.