FACES facial recognition led to a wrongful arrest — and it's not the first time
a lawsuit details how a 93% match score from the long-running FACES system led to the wrongful arrest of robert dillon, highlighting the dangers of treating algorithmic similarity as certainty in policing.
the aclu has filed a lawsuit against florida police departments after the FACES facial recognition system—a database containing tens of millions of images—generated a 93% similarity score that led to the wrongful arrest of robert dillon. despite dillon living 300 miles from the crime scene and having no prior connection to the area, officers treated the algorithm's match as a near-certain identification and allegedly failed to include exonerating evidence in warrant applications. the complaint highlights a fundamental misunderstanding of probability in policing, where a similarity score on facial features was conflated with the likelihood of identity. this incident serves as a critical warning for law enforcement agencies relying on legacy biometric tech, revealing how automation bias can override investigative due process and lead to severe civil rights violations.