AI-Assisted Random Selection for Citizen Budget Panels

What it is for

Eligibility & Public Benefits

Selects which volunteers are included in citizen panels for New York City's Participatory Budgeting Borough Assembly (The People's Money), determining who has the opportunity to directly influence public spending decisions.

NYC AI Use Cases Register — Panelot (2025)

Planning & Decision-making

Computes leximin-optimal probability distributions over possible panels to ensure fairness across the volunteer pool at the population level, supporting the CEC's resource allocation for participatory budgeting.

NYC AI Use Cases Register — Panelot (2025)

Run by: NYC Civic Engagement Commission
Where: No fixed location
Kept: No data retained
Shared with: Accountable organization
Your copy: You cannot see the data it holds about you. What you can do

What it is for

Eligibility & Public Benefits

Planning & Decision-making

Panelot is a non-profit tool used by New York City's Civic Engagement Commission to randomly select representative citizen panels for the Participatory Budgeting Borough Assembly process. It takes a pool of volunteers and quota requirements, then calculates a set of possible panels with assigned probabilities designed to maximize fairness to all volunteers. No personal identifying information is processed — only anonymized data about housing type, education, language, income, age, and zip code.

What it collects and what happens to it

Data taken in

Sensitive personal information

Pseudonymous data

Anonymized applicant records, each with a unique ID linking back to the CEC's internal list. Fields include: type of housing, education, language, income, age, and zip code. No direct identifiers (name, address, contact info) are present in the input processed by Panelot.

NYC AI Use Cases Register — Panelot (2025)

Operational data

Anonymized data

The desired panel size and quota constraints (e.g., the number of women in a panel of 100 must be between 47 and 53) that define the representativeness requirements for the selection process.

NYC AI Use Cases Register — Panelot (2025)

Processing

Optimization

Computes a leximin-optimal distribution over quota-compliant panels: first maximizing the lowest volunteer selection probability, then the second lowest, and so on. This optimization approach ensures fairness to all volunteers under representativeness constraints.

NYC AI Use Cases Register — Panelot (2025)

What it does

Deciding (Analytical AI)

Human decides

Panelot ranks and scores possible panels by computing leximin-optimal probabilities from structured volunteer data and quota constraints. The CEC then reviews the output list and contacts applicants — a human makes the final outreach and participation decisions.

NYC AI Use Cases Register — Panelot (2025)

Outputs

A recommendation or prediction

Pseudonymous data

A list of quota-compliant panels, each with an assigned selection probability, plus backup volunteers for some quota categories. The CEC uses this advisory output to contact applicants — it is not a binding automated decision, as humans conduct outreach and manage participant availability.

NYC AI Use Cases Register — Panelot (2025)

Run by

NYC Civic Engagement Commission

The NYC Civic Engagement Commission (CEC) deploys Panelot to select participants for the Borough Assembly phase of its annual Participatory Budgeting process, The People's Money.

NYC AI Use Cases Register — Panelot (2025)

Built by

Panelot

Panelot is a not-for-profit organization that created and maintains the Panelot tool on a pro-bono basis. It is an off-the-shelf system supplied without charge to the CEC.

NYC AI Use Cases Register — Panelot (2025)

Kept for

No data retained

The register does not specify a retention policy for Panelot's input or output data. Based on the tool's design as a one-time-per-cycle selection process, data retention beyond the Borough Assembly selection event is not described.
Duration: Not specified in source

NYC AI Use Cases Register — Panelot (2025)

Shared with

Not available to me

The probabilistic output panels and individual selection probabilities computed by Panelot are not directly accessible to individual volunteers or members of the public.

NYC AI Use Cases Register — Panelot (2025)

Available to the accountable organization

The output panel lists and probabilities are available to the NYC Civic Engagement Commission, which uses them to conduct outreach to selected participants and manage backups.

NYC AI Use Cases Register — Panelot (2025)

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 Algorithmic TransparencyPanelot is an open, not-for-profit system. The selection methodology — leximin-optimal probability distributions over quota-compliant panels — is publicly documented at https://panelot.org/. Volunteers can understand the general logic used to determine their selection chances.
Right to Non-discriminationThe leximin-optimality criterion is specifically designed to maximize the selection probability of the least-likely-to-be-selected volunteer first, reducing structural disadvantage. Quota constraints ensure demographic representativeness across housing type, education, language, income, age, and zip code.
Right to Be Informed of AI UseVolunteers who apply to participate in the Participatory Budgeting Borough Assembly are aware that an algorithmic tool is used to make the panel selection. The CEC contacts selected applicants directly from the output list.

Risks and safeguards

Financial & business harmA volunteer could be unfairly excluded from participating in the budgeting process due to quota constraints or optimization artifacts, effectively limiting their civic influence.Safeguard: The leximin criterion explicitly maximizes fairness by prioritizing the lowest selection probabilities first. Quota ranges (not fixed targets) prevent over-rigid exclusion. Human CEC staff contact backup volunteers when a selected participant becomes unavailable.
Civil liberties harmDemographic data used for quota-matching (housing, education, language, income, age, zip code) could be misused to exclude or stigmatize groups if the system were repurposed.Safeguard: Input data is anonymized with unique IDs not directly linked to personal identity within the tool. The system is operated by a public commission for a transparent civic purpose with published methodology. The tool is not used for surveillance or enforcement.