Francisco Arteaga was wrongfully charged with armed robbery based on facial recognition technology that identified his mugshot as a match from surveillance footage, leading to nearly four years of pretrial detention before he pleaded guilty to secure his release.
In November 2019, a gunman robbed the Buenavista Multiservices store in West New York, New Jersey, stealing $8,950. West New York police submitted surveillance images to the New Jersey Regional Operations Intelligence Center for facial recognition analysis but found no matches. Detectives then sent the footage to the NYPD Real Time Crime Center, where facial recognition technology identified Francisco Arteaga's mugshot as a 'possible match' in December 2019. Two store employees confirmed Arteaga from a photo array as the robber. Arteaga, who had an alibi and claimed he had never been to New Jersey, was charged and held as a pretrial detainee for nearly four years. His defense attorney sought detailed information about the facial recognition technology, including error rates and algorithms, but prosecutors claimed they couldn't provide this information because the match was made in another state. In June 2023, a state appellate panel ruled that Arteaga deserved access to these materials, but prosecutors still couldn't provide the details. Rather than continue fighting from prison, Arteaga pleaded guilty in November 2023 to return to his family. The case highlights gaps in regulation and oversight of facial recognition technology used by law enforcement agencies.
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