234 incidents on record · 2026HeadlightsIncident reports by Ellie Harris · Melbourne
5 new this weekLibrary last updated 14 October 2026
← The incident library
HD-INC-199
Hiring and recruitment technology · United States · 2025 · The agency enforcing a law on automated hiring tools relied on complaints that were not reliably reaching it and found far fewer potential compliance problems than state auditors found in the same 22 companies

New York City's AI Hiring Law Enforcer Found One Problem Where Auditors Found 17

By Ellie Harris · Filed Audit period July 2023 to June 2025; enforcement of Local Law 144 began 5 July 2023

Alleged: New York City Department of Consumer and Worker Protection (DCWP) developed or deployed the AI system implicated in this incident. Details are drawn from public reports; parties are presumed innocent of any wrongdoing not established by an official finding.

New York City's AI Hiring Law Enforcer Found One Problem Where Auditors Found 17

What happened

The New York State Comptroller reported in December 2025 that New York City’s agency responsible for enforcing its automated hiring law was missing potential breaches. Complaints were not reliably reaching the department, and state auditors found problems that its own reviews had missed.

New York City requires employers using certain automated hiring tools to check them for bias, publish the results and tell candidates when the tools will be used. The Department of Consumer and Worker Protection (DCWP) is responsible for enforcing those rules.

The Comptroller reported that when auditors called the city’s complaint line about automated hiring tools, most calls were sent to the wrong department or directed back to the employer. Very few reached DCWP.

The auditors also looked at employers whose websites DCWP had already reviewed. DCWP had identified one potential compliance problem. The state auditors reported finding at least 17 in the same group of employers.

Those findings were potential problems, not confirmed breaches of the law.

What an auditable version would have shown

The Comptroller reported that DCWP relied on complaints and company websites to find employers who might not be following the law. Its staff also lacked the technical expertise to assess automated hiring tools and had not sought help from the city’s technology office.

There was a problem with relying on complaints. If an employer did not tell a candidate that an automated tool had been used, the candidate might never know there was something to complain about.

DCWP needed a record of the complaints it received, where they went, which employers it reviewed and what it checked. That would have shown whether complaints were reaching the right people and whether its reviews covered the requirements of the law.

The auditors reported finding potential problems that DCWP had missed. A record of how each review was carried out would have helped establish why the department had reached different findings.

Where the gap was

The Comptroller reported that complaints about automated hiring tools were not reliably reaching DCWP. It also found that the department had identified far fewer potential compliance problems than the state auditors found when they examined the same employers.

The law requires employers to tell candidates when covered automated hiring tools will be used. But if an employer fails to give that notice, a candidate may have no idea that software played a part in assessing their application.

DCWP was relying on complaints from people who might not even know an automated hiring tool had been used.

A ConductRecord keeps a record of each decision an automated hiring tool contributes to, including the information used and the result.

A MetricRecord is designed to turn those records into signed aggregate figures, such as selection rates across different groups. An employer or auditor could use those figures to check whether a published bias audit reflects the decisions made using the tool.

ConductRecord and MetricRecord are Headlights designs. They could help employers show how their tools were used, but they would not fix the problems the Comptroller found with DCWP’s complaint handling and enforcement.

What governance should have looked like

The Comptroller reported that DCWP needed to improve how it received complaints, check published bias audits against the law’s requirements and look for potential breaches without waiting for someone to complain. The auditors also recommended that the department seek help from the city’s technology office.

DCWP agreed to implement or partially implement most of the recommendations. It rejected the recommendations to investigate why it had received so few complaints and to look for potential breaches through research and other leads. The department said that educating employers and candidates, alongside complaint-based enforcement, was the most efficient use of its resources.

Where an agency enforces rules on automated hiring tools, best practice would be to check that complaints reach the right department and regularly review whether employers are meeting the law’s requirements.

The agency should keep a record of which employers it reviewed, what it checked, what problems it found and what happened next. It should also be able to call on technical expertise when it needs to understand how a hiring tool works.

The Comptroller’s findings and DCWP’s response are public. The sources reviewed do not show whether the department completed the changes it agreed to make or whether a later review found that enforcement had improved.

Failure Pattern: the agency enforcing New York City’s law on automated hiring tools relied on complaints that were not reliably reaching it, and state auditors found at least 17 potential compliance problems in the same 22 companies where the agency had found one.

Governance Principle: a body that enforces rules on automated hiring tools should be able to show what it checked, against which requirement, and what it found, in a record that others can test.

The reference implementation of ConductRecord and MetricRecord is open source. It lives at github.com/saffronandindia/headlights-oss, Apache 2.0 licensed and free to install. The repository is public now.

Sources

The mailing list

Fresh incident reports every week. One email to match.

We add new incidents to the library regularly, and send a single short email each week with what's new. The library stays free and open; this is just how you keep up with it.

Sent through Substack. Unsubscribe in one click.

The record

An auditable system would have produced a signed, tamper-evident record the moment this happened: what the system did, the version that did it, the basis it acted on, and the action taken, and New York City Department of Consumer and Worker Protection (DCWP) could have produced it on demand.

This is the record the system as deployed did not produce in a signed, auditable form.

What this teaches
Capture what happened when it happens
What the system did, the version that did it, the basis it acted on, and the action taken, recorded at the moment, not reconstructed after.
Sign it, so no one has to trust the record-keeper
A tamper-evident entry. Edit it later and the signature breaks. The record does not ask for the benefit of the doubt.
Make it verifiable by anyone
A court, a regulator, a customer's lawyer can check the record themselves, without taking the company, or us, at our word.

Headlights summarises publicly reported AI incidents. All summaries are independently written, attributed to their original sources, and intended for research and educational purposes. Allegations are identified as such until established through official findings.

Last reviewed September 2026. This report is based on the sources listed above and reflects information available at the time of review; later developments may not be captured. Where a person is described as charged with or alleged to have done something, that allegation is unproven unless a conviction or a court or regulatory finding is stated. Headlights publishes journalism and commentary, not legal advice.

Want to write back?

Direct to my inbox.

ellie@useheadlights.com →