AI for CHROs · Independent decision intelligenceSource-backed reporting · No paid editorial rankings
CHRO AI Current

A workplace-AI publication for people leaders balancing workforce capability, employee experience, operational evidence, and the legal and human consequences of algorithmic decisions.

Home / Workforce Updates / NYC AEDT use needs a version-matched audit and notice record
Workforce updates

NYC AEDT use needs a version-matched audit and notice record

New York City's public guidance ties use of an automated employment decision tool to a bias audit within one year, a public summary, and notices. A CHRO still needs to prove that the audited tool, employer configuration, job population, decision use, and candidate-facing notice match what is actually operating on the decision date. A vendor name or generic audit report is not enough.

By CHRO AI Current Research DeskSource-backed analysis

Answer capsule

New York City's public guidance ties use of an automated employment decision tool to a bias audit within one year, a public summary, and notices. A CHRO still needs to prove that the audited tool, employer configuration, job population, decision use, and candidate-facing notice match what is actually operating on the decision date. A vendor name or generic audit report is not enough.

What the source establishes

  • NYC DCWP says Local Law 144 prohibits an employer or employment agency from using an AEDT unless it has been subject to a bias audit within one year of use.
  • The city page says information about the bias audit must be publicly available and required notices must be provided to employees or job candidates.
  • DCWP's page notes that the notice must be provided 10 business days before use of an AEDT and links the law, rule, FAQ, complaint route, and educational materials.
  • The overview does not establish whether a buyer's particular tool and use are covered, whether an audit matches the deployed version and population, or whether any workflow is fair, accessible, accurate, or lawful.

Classify the actual employment use

Inventory every automated component that scores, ranks, recommends, filters, prioritizes, summarizes, or otherwise materially influences a hiring or promotion decision. Record the job, location, candidate or employee population, decision stage, input data, output, human decision maker, vendor and product, model and configuration, integrations, dates, and whether the output can be ignored or corrected in practice. Do not assume a tool is outside scope because it is described as an assistant, embedded feature, general platform, or human-reviewed recommendation. Qualified employment counsel should determine legal applicability from the actual workflow; HR must supply accurate facts.

Bind the audit to the deployed version and use date

Create a version-matching record that links the audit report and public summary to the product release, configured features, selection criteria, data fields, covered job categories, employers or agencies, historical period, demographic categories, exclusions, calculations, auditor, publication date, and intended use. Compare that record with production before every covered use. A current-looking report can be mismatched if a model, resume parser, assessment, threshold, ranking rule, integration, job family, or candidate population changed. Set an expiry and a change trigger, and stop or narrow use when equivalence cannot be demonstrated.

Make notice and challenge operational

Map the candidate journey backward from the earliest tool use so the required notice is delivered with enough lead time and in an accessible form. Preserve the content, language, channel, delivery evidence, date, recipient, tool and job description, data information, contact route, and any accommodation or alternative-process handling. Train recruiters and agencies not to route a request back through the same automated judgment. Keep the public audit summary reachable, record questions and complaints, and give an accountable person authority to pause the workflow when notice, access, or version evidence is incomplete.

Govern changes beyond the annual calendar

Review releases continuously rather than treating the one-year audit window as permission for all intervening changes. Require vendors and internal owners to disclose model, feature, data, criterion, threshold, and integration changes before production. Re-run applicability, version matching, accessibility, performance across groups, job-related validation, notice, and human-review tests when the decision context changes. Monitor selection outcomes, overrides, correction requests, accommodations, missing data, complaints, and downstream employment effects without reducing people to one metric. Compliance evidence is a floor; it does not prove nondiscrimination, validity, fairness, or a defensible individual decision.

Turn this source into a reviewable decision

For AI for CHROs, use this briefing as a dated decision record rather than a substitute for the source. Preserve Automated Employment Decision Tools (AEDT), the exact URL, the August 31, 2026 review date, the supported facts above, the editorial interpretation, the limitations, and any buyer-specific evidence. Link that record to the decisions most directly affected: Recruiting and candidate support; Performance and work allocation; People analytics and employee listening; Workplace AI governance and change. State whether the source changes the scope, evidence requirement, control, sequence, or only the language used to describe the decision.

Before action, name the accountable owner, affected population and workflow, exact offering or configuration, source data and rights, human decision point, exception and appeal path, complete cost, expected benefit, failure and stop conditions, retained evidence, and next review date. Keep official facts, provider statements, buyer observations, representative tests, measured outcomes, editorial inferences, and unknowns visibly separate. Reopen the record when the source, offer, model, integration, data, policy, population, responsible person, or measured result changes.

Limitations and unknowns

NYC DCWP is the primary public-administration source. Its page summarizes Local Law 144 requirements for certain AEDT use, including a bias audit within one year, public information, notice, and a complaint route, and links the law, rule, and FAQ. This briefing provides decision preparation, not legal advice. The overview does not determine coverage for a particular employer, agency, job, location, tool, version, configuration, or use; validate an audit's population or method; or establish job relatedness, accessibility, nondiscrimination, notice delivery, human-review quality, or lawful operation. Current legal materials, deployment and change records, audit and population evidence, candidate-facing notices and challenge logs, representative testing, and qualified HR, talent, employee-relations, labor, accessibility, privacy, security, analytics, procurement, and employment-law review control.

Decision test

Ask whether the source changes the decision itself, the evidence required, the implementation sequence, or only the language used to describe an existing capability. Record which claims are directly supported, which are provider statements, which require an independent test, and which remain unknown. A source-linked review should make uncertainty easier to see, not bury it inside a blended score.

Questions to take into review

  • Does the tool materially influence who advances?
  • What validated job criteria support the output?
  • Which data is job-related and known to employees?
  • How are context and accommodations represented?
  • Could a person be reidentified?
  • What was the stated collection purpose?
  • Which roles and tasks change?
  • How are workers or representatives involved?
The publication supports research and executive decision preparation. It does not provide legal, financial, accounting, employment, clinical, cybersecurity, investment, procurement, or implementation advice.