Skip to content
This is NOT an official site of the City of New York. Click here for official AI reports.

NYC Office of Technology and Innovation

NYC - DTPR AI Register

Back to your search

AI-Assisted Facial Recognition for Criminal Investigations

What it is for

Safety & Security
  • The system is used to investigate criminal activity and to identify deceased and missing persons, enhancing law enforcement's ability to resolve cases.
Enforcement
  • Facial recognition is applied in the context of law enforcement investigations to help identify unknown subjects depicted in images obtained during those investigations.
Run by
New York Police Department (NYPD)
Where
No fixed location
Kept
Not stated by the NYC Office of Technology and Innovation.
Shared with
Accountable organization
Your copy
You cannot see the data it holds about you. What you can do

What it is for

Safety & SecurityEnforcement

The New York Police Department uses facial recognition software to compare images of unknown individuals obtained during investigations against a repository of lawfully held arrest and parole photos. The system generates a ranked list of possible matches for review by trained investigators — it does not make final identifications on its own. Members of the public whose photos appear in NYPD arrest or parole records may have their images searched without their direct knowledge.

What it collects and what happens to it

Data taken in

Biometric
Identifiable data
  • A still image depicting the face of an unknown individual obtained during a law enforcement investigation, submitted as a probe image for facial recognition analysis.
Sensitive personal information
Identifiable data
  • A controlled repository of lawfully obtained arrest and parole photographs against which the probe image is compared. These photos are linked to known individuals in NYPD records.

Processing

Biometric Recognition
  • The system applies facial recognition algorithms to match probe images of unknown individuals against a repository of arrest and parole photos, producing a ranked list of candidate matches. Training data is proprietary to the vendor DataWorks.

What it does

Deciding (Analytical AI)
Human decides
  • The system compares a probe image against a photo repository and returns a ranked pool of possible match candidates. Final identification decisions are made by trained Facial Identification Section investigators, not by the software alone.
Sensing (Perceptive AI)
Human decides
  • The system processes still images to detect and extract facial features from a probe image for subsequent comparison. It converts raw pixel data into structured facial representations.

Outputs

A recommendation or prediction
Identifiable data
  • A ranked pool of possible match candidates generated by the facial recognition software for review by trained Facial Identification Section investigators. The output is advisory — investigators must conduct further human analysis before any identification is acted upon.

Run by

New York Police Department (NYPD)
  • The NYPD deploys and operates this facial recognition system through its Facial Identification Section, using it to support criminal investigations, identify deceased persons, and locate missing persons.

NYC Algorithmic Tools Compliance Report — NYPD Facial Recognition Technology

Built by

DataWorks Plus
  • DataWorks is the commercial vendor that develops and maintains the facial recognition software under a procurement and professional services arrangement. Training data is proprietary to the vendor.

NYC Algorithmic Tools Compliance Report — NYPD Facial Recognition Technology

Kept for

Not stated by the NYC Office of Technology and Innovation.

Shared with

Not available to me
  • Members of the public cannot access the candidate match results, probe images, or the photo repository used by this system. The output is restricted to trained NYPD Facial Identification Section investigators.
Available to the accountable organization
  • Candidate match results and probe images are available to trained NYPD Facial Identification Section investigators for use in ongoing law enforcement investigations.

Stored

Not stated by the NYC Office of Technology and Innovation.

How to read the colours

Can it identify you?

Anonymized data
Data about people with the link to who is broken. Stripped of identifiers, blurred, aggregated, or noised so this system can’t reasonably tie a record back to an individual.
Pseudonymous data
Each person’s data is tied to a token (hash, ID, template) that lets this system recognise the same person across events, but the token itself doesn’t reveal a name. Reidentification is possible with extra information.
Identifiable data
The data either contains a direct identifier (name, address, account name, recognisable face or voice, plate number) or carries a token this system uses to look up legal identity during processing.

Who completes the loop?

Human decides
This mode suggests; a person decides what to do next. The AI is always advisory — a human is in the loop on every decision. Example: a triage tool ranks cases for a clinician who chooses which to see first.
Human executes
This mode decides; a person carries out the result. Example: an optimizer plans the day’s trash-collection routes, and drivers run them.
Autonomous
This mode decides and acts on its own. No person reviews each decision or carries out the resulting action.

Definitions from the DTPR standard. Amber is about your data, violet about who decides. The fuller the shape and the deeper the colour, the more identifying the data or the less a person is involved.

What you can do

Ask about this system

Questions go to the NYC Office of Technology and Innovation, not the vendor.

Your rights

  • Right to Be Informed of AI UseNew York City Local Law 49 of 2020 (the AEDT and algorithmic tools law) and the Algorithmic Transparency and Accountability requirement compel the NYPD to disclose the use of this tool publicly via the Algorithmic Tools Compliance Report. Individuals whose images are searched are generally not notified at the time of a search. Contact information for the NYPD's Facial Identification Section or a formal complaints channel is not specified in the register entry.
  • Right to Algorithmic TransparencyThe NYPD is required to publish this algorithmic tool disclosure under NYC's algorithmic accountability framework. The register entry describes the tool's general logic (probe image comparison against a photo repository) and autonomy level (informative, with human review). Vendor training data is described as proprietary, limiting full algorithmic transparency. The public may access the disclosure at the NYC Open Data portal.

Risks and safeguards

  • Civil liberties harmfacial recognition of individuals in law enforcement contexts can infringe on rights to privacy, freedom of movement, and due process, and documented accuracy disparities across demographic groups may cause misidentification — particularly affecting communities of color.Safeguard: the register states that the tool is used in combination with human analysis and additional investigation, and that investigators in the dedicated Facial Identification Section review all candidate matches before any action is taken. The system is described as informative rather than determinative.
  • Reputational harmfalse positive matches can cause innocent individuals to be wrongly associated with criminal investigations, with lasting reputational damage.Safeguard: the register states that all candidate matches are reviewed by trained Facial Identification Section investigators who apply human judgment and conduct additional investigation before any identification is used; however, no explicit error-rate thresholds or redress procedures are disclosed in the available register entry.