NJ Court Says Defendant Entitled to Detailed Discovery on the Facial Recognition Search that Identified Him
June 7, 2023
The Superior Court of New Jersey (New Jersey’s appellate court), ruled today in New Jersey v. Arteaga that a defendant is entitled to detailed information on how he was identified by a facial recognition search, even though the search itself was performed by an out-of-state agency. The court reversed-and-remanded a lower court decision denying discovery on most aspects of the NYPD’s facial recognition system and how it was used to identify Mr. Arteaga. You can read the opinion here.
This case arises out of the arrest and prosecution of Francisco Arteaga for an armed robbery. He was identified as a suspect based on a facial recognition search from the NYPD, conducted at the request of New Jersey police after their own facial recognition system returned no matches for the search of a poor-quality surveillance photo. Mr. Arteaga’s attorneys requested discovery on the provider and source code of the facial recognition system, how the system performed in NIST testing, known error rates, the photo actually submitted that identified Mr. Arteaga and other information they intend to use to to challenge the validity of Mr. Arteaga’s identification and arrest. The trial court denied Mr. Arteaga’s request, so many details of the facial recognition system would remain secret unless the appeals court disagreed.
EPIC filed an amicus brief with the Electronic Frontier Foundation and the National Association of Criminal Defense Lawyers arguing that discovery is necessary in cases like Mr. Arteaga’s because the risk of misidentification from a facial recognition system is unique to each search and varies greatly based on the system used, the database searched, the quality of the photograph submitted, and the demographics of the individual potentially misidentified. The brief explains that humans and facial recognition systems tend to make the same kinds of mistakes. Therefore, subsequent forms of human review, including photo lineups, are likely to reinforce instead of correct errors made by facial recognition systems. Discovery represents a last chance to correct misidentifications by facial recognition systems before innocent individuals are wrongfully convicted of crimes.
The court’s opinion relied on EPIC’s amicus stating “… defendant through his expert, and the secondary sources cited by defense counsel and amici, provide us convincing evidence of FRT’s novelty, the human agency involved in generating images, and the fact FRT’s veracity has not been tested or found reliable on an evidential basis by any New Jersey court.”