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 Assistant for Youth and Community Program Questions

What it is for

Inform
  • The chatbot answers commonly asked questions from the public about DYCD programs, providing predefined responses drawn from a staff-maintained knowledge base.

NYC Automated Decision Systems — DYCD Virtual Assistant Chatbot

Run by
Department of Youth and Community Development (DYCD)
Where
No fixed location
Kept
Not stated by the NYC Office of Technology and Innovation.
Shared with
Accountable organization, Vendor

What it is for

Inform

This chatbot helps members of the public get quick answers to common questions about programs offered by the NYC Department of Youth and Community Development (DYCD). It uses AI to understand the question and match it to a predefined answer from a staff-maintained knowledge base. The system operates fully automatically — no human reviews each response before it is sent. Users who need help with more complex questions are directed to the DYCD Community Connect team.

What it collects and what happens to it

Data taken in

About behaviour
Anonymized data
  • Free-text questions typed into the chatbot by members of the public, expressing their queries about DYCD programs. The register does not indicate that inputs are stored against a named individual.

NYC Automated Decision Systems — DYCD Virtual Assistant Chatbot

Operational data
Anonymized data
  • Staff-maintained knowledge base containing program information, curated question-answer pairs, and intent definitions used to train and operate the CLU and CQA models.

NYC Automated Decision Systems — DYCD Virtual Assistant Chatbot

Processing

Search & Retrieval
  • The Custom Question Answering (CQA) model searches the knowledge base for candidate answers, ranks them, and returns the best match to the user's inferred intent.

NYC Automated Decision Systems — DYCD Virtual Assistant Chatbot

What it does

Understanding (Semantic AI)
Autonomous
  • The Conversational Language Understanding (CLU) model captures the user's intent from free-text input, then directs the query to the Custom Question Answering (CQA) model to retrieve the best-matching predefined answer from the knowledge base. No human reviews each response before it is returned.

NYC Automated Decision Systems — DYCD Virtual Assistant Chatbot

Outputs

Generated content
Anonymized data
  • Predefined text answers selected and returned by the CQA model in response to the user's question. The content is drawn from the knowledge base rather than generated from scratch, but it is surfaced dynamically based on the matched intent.

NYC Automated Decision Systems — DYCD Virtual Assistant Chatbot

Run by

Department of Youth and Community Development (DYCD)
  • DYCD deploys and operates this chatbot to support its Community Connect team by providing the public with predefined answers to commonly asked questions about DYCD programs.

NYC Automated Decision Systems — DYCD Virtual Assistant Chatbot

Built by

Microsoft
  • Microsoft provides the underlying Azure Language cloud service — including Custom Question Answering and Conversational Language Understanding components — that powers the chatbot.

NYC Automated Decision Systems — DYCD Virtual Assistant Chatbot

Kept for

Not stated by the NYC Office of Technology and Innovation.

Shared with

Available to the accountable organization
  • DYCD's Information Technology Application Development team and the Community Connect team have access to system outputs and usage data (approximately 2,100 requests per day) to monitor performance and maintain the knowledge base.

NYC Automated Decision Systems — DYCD Virtual Assistant Chatbot

Available to vendor
  • Microsoft Azure processes the input questions and knowledge base content as part of providing the cloud-based language service. The extent of Microsoft's data access beyond service processing is not specified in the register.

NYC Automated Decision Systems — DYCD Virtual Assistant Chatbot

Stored

Stored on 3rd Party Cloud
  • The knowledge base and language model components are hosted on Microsoft Azure cloud infrastructure.
  • Duration: Not specified in the register

NYC Automated Decision Systems — DYCD Virtual Assistant Chatbot

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 UseUsers interacting with the DYCD website chatbot should be able to tell they are communicating with an automated system, not a human agent. The DYCD Community Connect team is available for questions beyond the chatbot's scope.
  • Right to Algorithmic TransparencyThe system uses Microsoft Azure Language's Conversational Language Understanding and Custom Question Answering components to match questions to predefined answers in a staff-maintained knowledge base. Information about this system is disclosed in the NYC AI use case register at https://www.nyc.gov/dycd.

Risks and safeguards

  • Psychological harmUsers with complex or nuanced needs may receive an inadequate or irrelevant predefined response, causing frustration or confusion when seeking help with important community services.Safeguard: The system is designed to redirect users with complex questions to the human Community Connect team, limiting the chatbot's role to basic informational queries.
  • Reputational harmThe system could return inaccurate or outdated program information to the public, potentially directing people to the wrong programs or giving incorrect eligibility details, which could harm DYCD's credibility and users' trust.Safeguard: Staff actively maintain and update the knowledge base; the 2025 upgrade to Microsoft Azure Language improves intent and answer matching accuracy.