AI-Assisted Language Interpreter Allocation for Poll Sites
What it is for
Eligibility & Public Benefits
- Determines which program-eligible languages receive interpretation services at which poll sites, and how many poll sites each language community is served at, based on proportional share of limited English proficient citizens of voting age (CVALEP).
Planning & Decision-making
- Supports the CEC's resource allocation planning by ranking program-eligible languages in order of magnitude of CVALEP population and distributing poll site assignments to balance inclusion of diverse language communities with fair access for each community's eligible voters.
- Run by
- NYC Civic Engagement Commission
- Where
- No fixed location
- Kept
- Retained Not specified; methodology updated with each new ACS five-year estimate cycle
- Shared with
- Accountable organization
- Your copy
- You cannot see the data it holds about you. What you can do
What it is for
This system uses U.S. Census data and a ranking algorithm to determine which languages receive interpretation services at which New York City poll sites during elections. It estimates the number of limited English proficient (LEP) citizens of voting age near each poll site and distributes interpreter resources proportionally across eligible languages. No personal identifying information about voters is collected or used. The system directly affects which language communities receive in-person interpretation assistance when casting their votes.
What it collects and what happens to it
Data taken in
About a measurement
Anonymized data
- American Community Survey (ACS) 2019–2023 five-year estimates providing citywide counts and Census Tract-level Public Use Microdata Samples (PUMS) of citizens of voting age with limited English proficiency, aggregated by language spoken and geographic area. No individual-level records are used.
About a place
Anonymized data
- Board of Elections data on the location of election districts and poll sites, used to determine which CVALEP populations are associated with each poll site catchment area via Thiessen polygon / Voronoi diagram construction.
Processing
Optimization
- Thiessen polygon / Voronoi diagram method assigns each residential census tract to its nearest poll site to estimate per-poll-site CVALEP populations. A descending-order greedy allocation then assigns poll sites to each language until that language's proportional share of total resources is met, with a guaranteed minimum of one poll site per eligible language.
Classification & Prediction
- Estimates the number of CVALEP citizens of voting age for each program-eligible language who could report to each polling site, using ACS five-year PUMS data and census tract geography. These estimates are the basis for ranking poll sites for each language in descending order of CVALEP concentration.
What it does
Deciding (Analytical AI)
Human decides
- The system ranks program-eligible languages by estimated CVALEP population size and scores poll sites by concentration of CVALEP speakers, producing a ranked allocation that human decision-makers at the CEC review before finalising interpreter deployments.
Outputs
A recommendation or prediction
Anonymized data
- A ranked list assigning each program-eligible language to a set of poll sites, and estimating CVALEP population per poll site per language. This is advisory — the CEC uses it to deploy interpreters but retains human authority over final service decisions. Output is at the language-community and poll-site level; no individual voter data is produced.
Run by
NYC Civic Engagement Commission
- The Civic Engagement Commission (CEC), under the Mayor's Office, is responsible for deploying and operating this ranking methodology to allocate language interpretation services at poll sites across New York City.
Built by
Not stated by the NYC Office of Technology and Innovation.
Kept for
Retained Not specified; methodology updated with each new ACS five-year estimate cycle
- The register does not specify a formal retention period for outputs. The methodology is updated periodically when new ACS data is released; the 2025 update used ACS 2019–2023 five-year estimates replacing prior ACS data.
- Duration: Not specified; methodology updated with each new ACS five-year estimate cycle
Shared with
Not available to me
- No personal data about individual voters is collected or disclosed. Outputs are aggregate population estimates and service-allocation plans at the language-community and poll-site level, not accessible by or about specific individuals.
Available to the accountable organization
- The CEC has access to the ranked language-to-poll-site allocation output for planning interpreter deployments. The Board of Elections also provides input data on poll site locations.
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.
- AI registerNYC Algorithmic Tools Compliance Report — Methodology for Poll Site Language AssistanceNYC Mayor's Office / Civic Engagement Commission, 2025 reporting year.
- AI registerNYC Algorithmic Tools Compliance Report
- AI registerNYC Algorithmic Tools Compliance Report
- AI registerNYC Algorithmic Tools Compliance Report
- AI registerNYC Algorithmic Tools Compliance Report
- AI registerNYC Algorithmic Tools Compliance Report
- AI registerNYC Algorithmic Tools Compliance Report
- AI registerNYC Algorithmic Tools Compliance Report
- AI registerNYC Algorithmic Tools Compliance Report
- AI registerNYC Algorithmic Tools Compliance Report
- AI registerNYC Algorithmic Tools Compliance Report
- AI registerNYC Algorithmic Tools Compliance Report
- AI registerNYC Algorithmic Tools Compliance Report
- AI registerNYC Algorithmic Tools Compliance Report
- AI registerNYC Algorithmic Tools Compliance Report
- AI registerNYC Algorithmic Tools Compliance Report
- AI registerNYC Algorithmic Tools Compliance Report
- Register entryPublished by the NYC Office of Technology and Innovation. Reference 3dec2406. This disclosure was drafted with AI assistance.Schema: ai@2026-05-06-beta
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 TransparencyThe methodology used to allocate poll site language services is publicly disclosed through the NYC Algorithmic Tools Compliance Report, available at the NYC Open Data portal. New Yorkers can review the logic of the ranking and proportional allocation approach in the published register entry.
- Right to Purpose LimitationThe system's inputs (Census CVALEP estimates and Board of Elections poll site data) are used solely to determine language interpretation service allocations for the CEC's poll site program. The register states that no identifying information is collected or disclosed, and data use is limited to this resource-allocation purpose.
Risks and safeguards
- Civil liberties harmSystematic under-service of smaller language communities, chilling participation for LEP voters whose language is not allocated a poll site, or exclusion of communities due to proxy measure limitations.Safeguard: Charter-mandated minimum of one poll site per eligible language; proportionality formula balancing both community inclusion and total-population equity; public disclosure of methodology; annual data updates using the most recent ACS five-year estimates; explicit exclusion of Board of Elections-served languages from CVALEP counts to avoid double-counting.
- Reputational harmA language community could be publicly ranked low or excluded from services based on proxy CVALEP estimates that imperfectly capture actual LEP registered voter counts, potentially stigmatizing that community.Safeguard: The methodology explicitly acknowledges the proxy limitation ('no dataset is currently available that reliably captures the number of LEP registered voters') and frames rankings as estimates. Transparency via public register disclosure allows communities to scrutinize and contest allocations.