New York City Police Department

The NYPD's facial recognition impact and use policy says the department compares 'probe images' with a limited photo repository of arrest and parole photographs to identify suspects recorded on camera at robberies, burglaries, assaults, shootings and other crimes, and to identify people who cannot identify themselves. The policy names Dataworks Plus and Idemia as the manufacturers and says only algorithms evaluated by the National Institute of Standards and Technology are used. It states that a possible match is only an investigative lead and that 'No one has ever been arrested solely based on facial recognition results by the NYPD.' It also states that 'The NYPD does not perform real-time facial recognition' and that department systems cannot conduct real-time facial recognition. Access is limited to NYPD facial recognition investigators. The policy's revision summary records that a statement that the facial recognition software does not use artificial intelligence and machine learning was removed. The policy says the NYPD has used facial recognition since 2011; the New York City Office of Technology and Innovation's algorithmic tools report for 2024 lists the NYPD's facial recognition technology as first used in October 2011 and names Dataworks as the vendor. At the City Council Committee on Public Safety hearing on 19 February 2025, according to an automated transcript on citymeetings.nyc, Deputy Commissioner Michael Gerber said the three tools the NYPD reports on annually are Patternizr, ShotSpotter and facial recognition, that 'these are technologies that use AI in some form', and that 'This is not AI predictive policing'.