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

AI-Assisted Drafting of City Agencies Reference Dataset

What it is for

Planning & Decision-making
  • The tool was used to create a standardized, analysis-ready, machine-readable reference of New York City agencies and other organizations with NYC-specific governance functions, supporting both public discovery via NYC.gov and internal analytics.

NYC AI Use Cases

Run by
Office of Technology and Innovation (OTI)
Where
No fixed location
Kept
Retained indefinitely (ongoing dataset)
Shared with
Accountable organization, Download

What it is for

Planning & Decision-making

This tool used a large language model (Claude by Anthropic) to help build and maintain a standardized reference list of New York City agencies and governance organizations. The model assisted with setting naming rules, deduplicating records, and transforming raw source lists into a clean, machine-readable dataset. Human reviewers checked all outputs before they were published to NYC Open Data and the NYC.gov Agency Directory. The AI operated in an advisory capacity — all changes required structured human quality assurance before publication.

What it collects and what happens to it

Data taken in

Operational data
Anonymized data
  • Internal and external sources on New York City agencies and governance organizations including the Citywide Org Chart, the Mayor's Office of Operations' list, and lists from the Department of Records and Information Services and the Department for Citywide Administrative Services.

NYC AI Use Cases

Processing

Language Models
  • Claude by Anthropic (off-the-shelf LLM) was used to ingest, normalize, deduplicate, validate, and publish agency records, with Unicode/encoding fixes, name parsing, and auditable transformations at each step of the pipeline.

NYC AI Use Cases

What it does

Understanding (Semantic AI)
Human decides
  • The LLM was used to understand and normalize entity names, institute acceptance rules for inclusion and naming conventions, and support record linkage and deduplication across multiple source lists.

NYC AI Use Cases

Deciding (Analytical AI)
Human decides
  • The LLM classified and scored candidate agency records for inclusion and canonical naming in the reference dataset, with human reviewers making final acceptance decisions.

NYC AI Use Cases

Outputs

Operational data
Anonymized data
  • A single, analysis-ready table with preferred names, common alternates, leadership, and reporting lines that powers both public discovery (NYC.gov Agency Directory) and internal analytics.

NYC AI Use Cases

Run by

Office of Technology and Innovation (OTI)
  • The Office of Technology and Innovation's Office of Data Analytics deployed this tool to create and maintain the NYC Agencies and Governance Organizations reference dataset for public and internal use.

NYC AI Use Cases

Built by

Anthropic
  • Anthropic's Claude is the off-the-shelf large language model used to assist with data ingestion, normalization, deduplication, validation, and publication pipeline tasks. Training data is proprietary to Anthropic.

NYC AI Use Cases

Kept for

Retained indefinitely (ongoing dataset)
  • The output dataset is a living reference published to NYC Open Data. No explicit retention period is stated in the register entry; the dataset is maintained on an ongoing basis.
  • Duration: indefinitely (ongoing dataset)

NYC AI Use Cases

Shared with

Available to the accountable organization
  • The output dataset is available to the Office of Technology and Innovation and other NYC agencies for internal analytics purposes, as well as published publicly via NYC Open Data.

NYC AI Use Cases

Available to download
  • The resulting NYC Agencies and Governance Organizations dataset is publicly available for download on NYC Open Data at the source URL.

NYC AI Use Cases

Stored

Stored on 3rd Party Cloud
  • The output dataset is hosted on NYC Open Data, which is operated on cloud infrastructure. The LLM processing was performed via Anthropic's cloud-based Claude API; training data is proprietary to Anthropic.
  • Duration: indefinitely (ongoing dataset)

NYC AI Use Cases

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 Algorithmic TransparencyMembers of the public can review the published NYC Agencies and Governance Organizations dataset on NYC Open Data to understand what entities were included and how they were named. The register entry describes the general pipeline logic used (ingest, normalize, deduplicate, validate, publish). For further information about the methodology, contact the NYC Office of Technology and Innovation.

Risks and safeguards

  • Reputational harmThe LLM could misclassify, misspell, or incorrectly exclude/include agency names, resulting in an authoritative public reference that misrepresents official NYC governance structures.Safeguard: Structured human quality assurance (peer review, change logs, and schema validation) was applied before any scripted export to NYC Open Data; the dataset is also versioned and subject to ongoing review.