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 Fraud Investigations

What it is for

Enforcement
  • Supports DOI investigations of fraud and other criminal activity within its jurisdiction. Facial recognition outputs serve only as investigative leads — not as final identifications — requiring additional investigative steps.

NYC Algorithmic Tools Compliance Report — Facial Recognition Technology (DOI)

Run by
Department of Investigation (DOI)
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

Enforcement

The NYC Department of Investigation uses facial recognition software to compare images and videos gathered during investigations with lawfully held images, generating a pool of possible matches. A trained DOI examiner then manually reviews every potential match before any investigative action is taken — the system never makes a final identification on its own. The tool is used to support investigations into fraud and other criminal activity within DOI's jurisdiction.

What it collects and what happens to it

Data taken in

Biometric
Identifiable data
  • Images or videos of people and objects obtained during a DOI investigation. These inputs contain facial imagery and may directly identify individuals who are the subject of investigation.

NYC Algorithmic Tools Compliance Report — Facial Recognition Technology (DOI)

Sensitive personal information
Identifiable data
  • The vendor's model was trained on publicly available open-source media data depicting individuals. This training data constitutes personal information handled by the system.

NYC Algorithmic Tools Compliance Report — Facial Recognition Technology (DOI)

Processing

Biometric Recognition
  • The system performs face-matching: it compares uploaded investigative images against a database of lawfully possessed images to identify possible matches. It does not produce a verified identity — outputs are candidate matches for human review.

NYC Algorithmic Tools Compliance Report — Facial Recognition Technology (DOI)

What it does

Sensing (Perceptive AI)
Human decides
  • The system analyzes uploaded images or videos to detect and extract facial features, which are then compared against a database of lawfully possessed images. All outputs are reviewed by a trained DOI examiner before use.

NYC Algorithmic Tools Compliance Report — Facial Recognition Technology (DOI)

Deciding (Analytical AI)
Human decides
  • The system scores and ranks candidate images by similarity to the query image, generating a pool of possible matches. The Autonomy type is Informative — outputs are advisory and a human examiner decides whether any match is reliable.

NYC Algorithmic Tools Compliance Report — Facial Recognition Technology (DOI)

Outputs

Biometric
Identifiable data
  • Lawfully possessed images of possible matches, if found. The output is a ranked pool of candidate facial images drawn from lawfully held databases, presented to a human examiner for visual evaluation.

NYC Algorithmic Tools Compliance Report — Facial Recognition Technology (DOI)

A recommendation or prediction
Anonymized data
  • A pool of possible matches serves as an investigative lead only. The register explicitly states: 'A match serves as an investigative lead for additional investigative steps and does not constitute a positive identification.'

NYC Algorithmic Tools Compliance Report — Facial Recognition Technology (DOI)

Run by

Department of Investigation (DOI)
  • The NYC Department of Investigation deploys this facial recognition tool for use by its Compliance department to assist investigations of fraud and other criminal activity within its jurisdiction. Trained DOI examiners review all potential matches before any investigative lead is acted upon.

NYC Algorithmic Tools Compliance Report — Facial Recognition Technology (DOI)

Built by

Undisclosed Facial Recognition Vendor
  • The vendor name is not publicly disclosable. The tool is an off-the-shelf product and the vendors provide ongoing technical assistance. Confidentiality agreements are in place with the vendors.

NYC Algorithmic Tools Compliance Report — Facial Recognition Technology (DOI)

Kept for

Not stated by the NYC Office of Technology and Innovation.

Shared with

Available to the accountable organization
  • Outputs (possible match images and investigative leads) are available to DOI investigators and examiners for use in active investigations. Access is limited to DOI staff involved in the investigation.

NYC Algorithmic Tools Compliance Report — Facial Recognition Technology (DOI)

Not available to me
  • Individuals who are the subject of a DOI investigation involving facial recognition are not notified of or given access to the system's outputs. The register does not describe any individual access mechanism.

NYC Algorithmic Tools Compliance Report — Facial Recognition Technology (DOI)

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 UseMembers of the public whose images may be searched are not directly notified that facial recognition is in use, as this tool supports active investigations. Transparency is provided at the system level through the NYC Algorithmic Tools Compliance Report, publicly available at the source URL.
  • Right to a Human ReviewAll potential matches generated by the system are reviewed by a trained DOI examiner before any investigative action is taken. The system does not make final identifications autonomously — human review is mandatory for every output.
  • Right to Algorithmic TransparencyHigh-level information about this system is disclosed through the NYC Algorithmic Tools Compliance Report. The vendor name is not publicly disclosable due to confidentiality agreements. Detailed algorithmic logic and model architecture are not disclosed.

Risks and safeguards

  • Civil liberties harmFacial recognition may misidentify individuals, leading to wrongful targeting in criminal investigations and implicating civil liberties including freedom from arbitrary surveillance and due process.Safeguard: The system is constrained to an informative/advisory role — all outputs are reviewed by a trained DOI examiner before use as investigative leads, and a match 'does not constitute a positive identification.' Use is restricted to lawfully possessed images and governed by agency policy and applicable laws.
  • Reputational harmA false-positive facial match could stigmatize or falsely implicate an individual in a criminal investigation, causing reputational damage.Safeguard: Every potential match undergoes visual analysis and evaluation by a trained DOI examiner 'to assess reliability of a match consistent with agency policy and applicable laws,' and the output is explicitly categorized as an investigative lead, not an identification.