# NYC Local Law 144

NYC Local Law 144 regulates the use of automated employment decision tools in hiring and promotion, requiring bias audits and notice to candidates and employees. It took effect on July 5, 2023.

NYC Local Law 144 (LL 144) is a municipal ordinance in New York City that regulates the use of automated employment decision tools (AEDTs) in hiring and promotion decisions. The law, officially titled "Automated Employment Decision Tools," was enacted in December 2021 and took effect on July 5, 2023. It is among the first laws in the United States to specifically target algorithmic bias in employment, setting a precedent for similar legislation in other jurisdictions. The law applies to employers and employment agencies operating in New York City, regardless of their physical location, if they use AEDTs to screen candidates or employees for promotion.

The core requirement of LL 144 is that any AEDT used for employment decisions must undergo an independent bias audit within one year before its use. The audit must assess the tool's impact on race, ethnicity, and sex, and the results must be published on the employer's website. Additionally, employers must provide notice to candidates and employees that an AEDT will be used, along with information about the tool's characteristics and how to request an alternative selection process or accommodation. The law also prohibits the use of AEDTs that rely solely on a single factor, such as a personality test, unless that factor is job-related.

## Legislative Background and Intent

LL 144 was introduced by New York City Councilmember Laurie Cumbo in 2020, amid growing concerns about the use of [AI](https://www.wikiprompt.org/wiki/artificial-intelligence) in hiring. The law was part of a broader package of algorithmic accountability measures, including a separate law on automated decision systems in city agencies. The intent was to address documented cases where automated tools discriminated against protected groups, such as a 2018 study showing that an AI recruiting tool from a major tech company penalized resumes containing the word "women's." The law aimed to create transparency and accountability without banning the technology outright, instead mandating audits and disclosure.

The New York City Department of Consumer and Worker Protection (DCWP) was tasked with implementing the law. The DCWP issued proposed rules in 2022 and final rules in April 2023, clarifying key definitions and procedures. The effective date was pushed from January 1, 2023, to July 5, 2023, to give employers time to comply. The final rules defined an AEDT as any computational process derived from [machine learning](https://www.wikiprompt.org/wiki/machine-learning), statistical modeling, or other data-processing techniques that issues a simplified output, such as a score or classification, used to substantially assist or replace discretionary decision-making.

## Key Provisions and Requirements

The bias audit is the centerpiece of LL 144. It must be conducted by an independent auditor who is not the vendor or developer of the AEDT, nor an employee of the employer. The audit must use historical data from the employer's own use of the tool, or test data if historical data is unavailable. The audit must calculate the selection rate for each category of race, ethnicity, and sex, and compare it to the highest selection rate, using a metric similar to the four-fifths rule from the Equal Employment Opportunity Commission. If the selection rate for a group is less than four-fifths of the highest group's rate, the audit must flag a potential bias.

The audit results must be published on the employer's website, including the date of the audit, the tool's name and version, and the summary of results. The publication must occur within 30 days of the audit's completion. Employers must also retain audit records for at least three years. The law does not require a specific passing score, but if an audit reveals bias, the employer must either take corrective action or stop using the tool.

Notice requirements are equally important. Employers must provide a notice to each candidate or employee who is subject to an AEDT at least 10 business days before the tool is used. The notice must describe the tool's name, the type of data it collects, the tool's scoring or classification logic, and the employer's policy on alternative selection procedures. The notice must also explain how to request an accommodation, such as a manual review or a different assessment method. The law requires that the notice be accessible, including in languages other than English if the employer typically communicates in those languages.

## Scope and Definitions

LL 144 applies to "employment decisions" including hiring, promotion, and retention. It covers both initial screening and ongoing evaluation. The law defines an AEDT as a tool that "substantially assists or replaces discretionary decision-making." This means the tool must be used to make a decision that would otherwise require human judgment. A tool that merely provides raw data, such as a resume parser that extracts keywords without scoring, is not covered. However, a tool that ranks candidates or predicts job performance is covered.

The law applies to employers with at least one employee in New York City, as well as employment agencies that refer candidates to NYC employers. It does not apply to tools used solely for internal administrative purposes, such as scheduling or payroll. It also does not apply to tools that are entirely manual, such as a human reviewing a paper resume. The law has extraterritorial reach: an employer based in California that hires for a NYC position must comply if it uses an AEDT for that hiring.

## Enforcement and Penalties

Enforcement of LL 144 is civil, not criminal. The DCWP can investigate complaints and conduct its own audits. Penalties for first violations are $500 per day, and for subsequent violations, $1,500 per day. Each day of non-compliance is a separate violation. The law also creates a private right of action, allowing individuals to sue for injunctive relief, damages, and attorney's fees. However, the private right of action is limited to cases where the employer failed to publish an audit or provide notice, not for discriminatory outcomes per se.

The DCWP has published a complaint form and has stated it will prioritize cases involving clear violations, such as no audit or no notice. As of 2024, the DCWP had not publicly reported any enforcement actions, but the law has been cited in several private lawsuits. In one notable case, a job applicant sued a retail chain for using an AI video interview tool without providing notice, alleging violations of LL 144. The case was settled in 2024, but the terms were not disclosed.

## Impact and Industry Response

The law has had a significant impact on the hiring technology industry. Vendors of [ML](https://www.wikiprompt.org/wiki/machine-learning)-based screening tools, such as those using [neural networks](https://www.wikiprompt.org/wiki/neural-network) or [large language models](https://www.wikiprompt.org/wiki/large-language-model), have had to adapt their products to support bias audits. Many vendors now offer audit-ready features, such as automated calculation of selection rates and generation of audit reports. Some vendors have also begun to market their tools as "LL 144 compliant," although compliance is ultimately the employer's responsibility.

Employers have responded by either conducting audits internally (using third-party auditors) or by discontinuing use of certain AEDTs. A 2023 survey by a HR consulting firm found that 23% of NYC-based employers had stopped using at least one automated hiring tool due to LL 144. The cost of an independent bias audit ranges from $5,000 to $30,000 per tool, depending on data volume and complexity. This has been a barrier for small employers, leading some to rely on manual review instead.

The law has also influenced other jurisdictions. In 2023, the state of California passed a similar law (AB 331) that requires bias audits for automated decision systems, though it was vetoed by the governor. The city of Washington, D.C., and the state of Illinois have considered similar measures. The Equal Employment Opportunity Commission (EEOC) has cited LL 144 as a model for federal regulation, and in 2023, the EEOC issued technical guidance on algorithmic fairness that references the NYC law.

## Criticisms and Limitations

Critics have raised several concerns about LL 144. First, the law's focus on race, ethnicity, and sex does not cover other protected categories such as age, disability, or religion. Second, the four-fifths rule is a blunt instrument that may not capture subtle forms of bias, especially in small sample sizes. Third, the law does not require the audit to assess the tool's accuracy or job-relatedness, only its disparate impact. Fourth, the notice requirements are onerous and may not be practical for high-volume hiring, where candidates are screened in minutes.

Another limitation is the law's reliance on historical data. If an employer has not used the tool before, it must use test data, which may not reflect real-world conditions. The law also allows employers to use "alternative selection procedures" but does not define what those are, leaving room for evasion. Some legal scholars have argued that the law's private right of action is too narrow, as it does not allow plaintiffs to challenge the audit's methodology. Finally, the law does not require the audit to be made public in a machine-readable format, making it difficult for researchers to analyze.

## Future Developments

As of 2025, the DCWP has not announced any amendments to LL 144, but it has stated it will monitor the law's effectiveness. The agency has published guidance on best practices for audits and has held public hearings. There is ongoing debate about whether to extend the law to cover other types of automated decisions, such as performance monitoring or termination. Some advocates have called for a ban on certain AEDTs, such as those using [deep learning](https://www.wikiprompt.org/wiki/deep-learning) with opaque decision-making, but no such proposal has been formally introduced.

The law has also spurred academic research on algorithmic auditing. Researchers at [MIT CSAIL](https://www.wikiprompt.org/wiki/mit-csail) and [Stanford AI Lab](https://www.wikiprompt.org/wiki/stanford-ai-lab) have developed open-source tools for bias auditing that comply with LL 144's requirements. These tools use techniques such as [data augmentation](https://www.wikiprompt.org/wiki/data-augmentation) and [model pruning](https://www.wikiprompt.org/wiki/model-pruning) to simulate different demographic distributions. The field of algorithmic fairness, which includes researchers like [Samy Bengio](https://www.wikiprompt.org/wiki/samy-bengio) and [Anima Anandkumar](https://www.wikiprompt.org/wiki/anima-anandkumar), has increasingly focused on practical compliance frameworks.

In the broader context, LL 144 is part of a global trend toward regulating AI in employment. The European Union's AI Act, which was adopted in 2024, includes similar requirements for high-risk AI systems, including those used in hiring. The law has been cited in academic papers and industry reports as a pioneering example of "audit-based regulation." Whether it will be effective in reducing bias remains an open question, but it has certainly changed the conversation about how to govern [AI](https://www.wikiprompt.org/wiki/artificial-intelligence) in the workplace.

## Conclusion

NYC Local Law 144 represents a significant step in the regulation of automated decision-making in employment. By requiring bias audits and transparency, it aims to mitigate the risks of algorithmic discrimination while allowing the use of beneficial technologies. The law's implementation has been challenging, with debates over definitions, audit methodology, and enforcement. However, it has established a framework that other jurisdictions are likely to follow. As AI continues to evolve, the principles of LL 144 - independent audits, public disclosure, and candidate notice - may become standard practice worldwide. The law's ultimate success will depend on whether it actually reduces bias without unduly burdening employers or stifling innovation.

---
Source: https://www.wikiprompt.org/wiki/nyc-local-law-144
License: CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/)
Last updated: 2026-09-12T16:24:04.728785+00:00