Man sues Florida cops over arrest spurred by "93% match" in facial recognition

Key Points:

• Robert Dillon sued Florida police for wrongful arrest based on a faulty 93% facial recognition match from a low-quality photo of McDonald's surveillance footage, despite evidence showing he was not in the area at the time.
• The lawsuit alleges that officers concealed exculpatory evidence, including license plate data and Dillon's denial of involvement, and failed to pursue available investigative leads such as mobile ordering records and cell phone location data.
• Dillon was arrested and prosecuted for over two months for attempting to lure a child, a charge later dropped; he suffered reputational harm, financial strain, and emotional distress due to the wrongful arrest.
• The facial recognition system used, FACES, is criticized for its unreliability and lack of clear probabilistic meaning behind confidence scores, with the police disregarding their own policy that such results do not constitute probable cause.
• The lawsuit demands damages and reforms in police use of facial recognition technology and highlights concerns about the lead investigator's prior disciplinary issues and the broader risks of AI-driven law enforcement errors.