NYC Local Law 144: The Bias Audit and Notice Rules for Automated Hiring Tools

The City Council approved the bill on 10 November 2021 and it was enacted on 11 December 2021 as Local Law 144 of 2021. The text says it takes effect on 1 January 2023. The Department of Consumer and Worker Protection, known as DCWP, adopted an implementing rule and began enforcing both on 5 July 2023.

The law added a subchapter to the city's Administrative Code. Section 20-870 holds the definitions, 20-871 the requirements and 20-872 the penalties. DCWP's rule fills in the detail on what an audit must calculate and how notice can be given.

It regulates process. An employer has to audit, publish and notify. The law does not set a score a tool must reach, and DCWP's FAQ says it does not require any specific action based on the results of an audit.

Rows cite the law as published by the City Council, the DCWP rule as adopted, or DCWP's own pages. All were read on 12 November 2026.

The law and the rule use different numbering. Sections beginning 20 are in the Administrative Code, and sections beginning 5 are in DCWP's rule.

The law defines an automated employment decision tool as a computational process derived from machine learning, statistical modeling, data analytics or artificial intelligence that issues a simplified output, such as a score, classification or recommendation, used to substantially assist or replace discretionary decision making. Junk email filters, calculators and spreadsheets are excluded.

DCWP's rule narrows the key phrase to three situations. The employer relies solely on the simplified output with no other factors. Or the output is one of several criteria and is weighted more than any other. Or the output is used to overrule conclusions from other factors, including human decision-making.

On location, DCWP's FAQ says the law applies to jobs with at least part-time work in a New York City office and to fully remote jobs tied to a New York City office. It covers screening as well as final decisions. The FAQ says it does not cover using a tool to source candidates from resume banks or to reach out to people who have not applied.

Whether a specific product meets the three-part test depends on how an employer actually uses it. The same software can be covered at one company and not at another.

The rule defines the terms. A selection rate is the share of people in a category who move forward or are assigned a classification. The rule's example is 100 Hispanic women applying and 40 being selected for an interview, a selection rate of 40 percent. An impact ratio is a category's selection rate divided by the rate of the most selected category.

For tools that score people, the audit uses a scoring rate, which is the share of a category scoring above the sample's median. The impact ratio is then that rate divided by the rate of the highest-scoring category.

The audit must also state how many people were assessed but fall in an unknown category. DCWP's FAQ says categories making up less than 2 percent of the data may be excluded from the calculations. The rule sets no pass mark for an impact ratio.

The rule defines an independent auditor as a person or group capable of objective and impartial judgment. Someone is not independent if they were involved in using, developing or distributing the tool, if they work for the employer or the vendor during the audit, or if they have a direct or material indirect financial interest in either. DCWP's FAQ says the agency does not approve auditors.

The audit must use historical data, meaning data collected while the tool was used to assess candidates. The data can come from more than one employer. An employer may rely on an audit built on other employers' data only if it contributed its own data to the auditor or has never used the tool.

Test data is the fallback. An employer may rely on an audit that uses test data if there is not enough historical data for a statistically significant audit. The published summary must then explain why historical data was not used and how the test data was produced.

The audit expires. An employer may not use or keep using a tool if more than one year has passed since its most recent bias audit.

Before using a tool, the employer must post the date of the most recent audit, a summary of results and the distribution date, which is the date the employer began using that tool. The rule requires this on the employment section of the website in a clear and conspicuous manner. A clearly labeled link is acceptable. The summary must stay posted for at least six months after the tool is last used.

Notice goes to candidates and employees who live in the city. The law requires it at least 10 business days before use. It must say that an automated tool will be used and which job qualifications and characteristics it will assess, and it must explain how to request an alternative selection process or accommodation. The rule adds that nothing in it requires an employer to provide an alternative process.

• For candidates: notice on the employment section of the website, in the job posting, or by mail or email.
• For employees up for promotion: notice in a written policy, in the job posting, or by mail or email.
• In every case, at least 10 business days before the tool is used.
• On written request: the type of data collected, its source and the retention policy, within 30 days, if not already on the website.

Section 20-872 sets a civil penalty of not more than $500 for a first violation and for each additional violation on the same day. Each subsequent violation carries between $500 and $1,500.

The amounts add up through counting. Each day a tool is used in violation is a separate violation. Each failure to give a required notice to a candidate or employee is a separate violation as well. The law firm DLA Piper summarizes the exposure as up to $1,500 per violation per day.

The law also preserves other routes. It does not limit a candidate's right to bring a civil action or the authority of the city's Commission on Human Rights. A bias audit that shows a poor impact ratio is public, and anyone can read it.

The New York State Comptroller published audit 2024-N-6 on 2 December 2025, covering July 2023 through June 2025. It found that DCWP received two complaints about automated employment tools in that period, and that the complaint process did not ensure complaints made to the city's 311 line reached DCWP. DLA Piper reports that 75 percent of the auditors' test calls to 311 were misrouted.

The audit also tested DCWP's own reviews. DCWP had looked at 32 companies and identified one instance of non-compliance. The auditors reviewed the same companies and identified at least 17 instances of potential non-compliance. DCWP officials acknowledged they lacked the technical expertise to evaluate these tools.

DLA Piper, which describes the Comptroller as calling enforcement ineffective, reports that DCWP agreed to implement most of the recommendations. Those include routing complaints correctly, training staff, adopting written procedures and using interviews and tool demonstrations. No later enforcement figures were found for this article.

DLA Piper's advice after the state audit is to assume closer review. Its list starts with an inventory and ends with records. The points below follow that order and describe what the law and rule ask for.

The firm also suggests that an employer with an existing audit should not assume it meets the rule, given how many gaps the state auditors found in published summaries.

• List every third-party and in-house tool used in hiring and promotion, and note how its output is used in the decision.
• For each covered tool, confirm the date of the last bias audit and who performed it.
• Check that the summary and distribution date are posted clearly on the employment section of the website.
• Check that notice reaches city residents at least 10 business days before the tool is used.
• Keep dated records of the inventory, each audit, each posting and each notice.