What Local Law 144 Requires of New York City Employers

Local Law 144 of New York City requires employers to conduct bias audits on automated employment decision tools (AEDTs). These audits must be performed by qualified third parties and must evaluate whether the tools have a disparate impact on protected groups. The law applies to employers with four or more employees who use AEDTs for employment decisions such as hiring, promotion, or termination. The audit must be published annually or whenever there is a material change to the tool. The purpose of these audits is to ensure that employment practices do not unintentionally discriminate against individuals based on protected characteristics such as race, gender, or age.

Key Requirements of Local Law 144

Employers must commission an audit of any AEDT they use in employment decisions. The audit must be conducted by a qualified third-party auditor who is not employed by the company. The auditor must evaluate the tool’s potential for bias and must provide a written report. The report must include findings on whether the tool has a disparate impact on protected groups. The employer must make this report available to employees and prospective employees upon request. The audit must be completed within 120 days of commissioning, and the results must be published annually or whenever there is a significant change to the tool. The law also requires that employers maintain records of the audit for at least three years.

• Employers must commission an audit of any AEDT used in employment decisions
• The audit must be performed by a qualified third-party auditor
• The audit must evaluate potential disparate impact on protected groups
• The audit report must be published annually or upon material change
• Employers must make the report available upon request
• Records must be retained for at least three years

Implementation Examples

For example, a large retail company using an AEDT to screen job applicants must commission an audit through an independent auditor. The auditor evaluates whether the tool disproportionately rejects candidates from certain racial or gender groups. The company must then publish the audit results and make them accessible to job seekers. If the tool is updated or replaced, a new audit must be commissioned. The company must also keep records of the audit, including the auditor’s report, for at least three years. Another example involves a financial services firm that uses an AEDT to assess loan applicants. The firm must ensure that the tool does not have a disparate impact on applicants from protected groups. The audit must be completed and published, and any identified issues must be addressed through tool modifications or additional training.

Employers must also ensure that their AEDTs comply with broader AI governance frameworks. The EU AI Act, which became effective in February 2025, introduces obligations for AI systems, including those used in employment. The act prohibits certain high-risk AI practices and requires AI literacy for staff. Employers must ensure that their AEDTs meet these requirements, particularly regarding transparency and accountability. The Digital Omnibus on AI, which came into force in July 2026, further clarifies these obligations. Companies must ensure that any generative AI systems placed on the market before August 2026 are marked as machine-readable by December 2026. These obligations apply to any AI tool used in employment decisions, including those that are not directly linked to hiring but affect promotion or performance reviews.

Organisations must also consider the ISO/IEC 42001:2023 standard for AI management systems. This standard provides a framework for implementing AI governance, including risk management, transparency, and accountability. Companies that adopt this standard can demonstrate due diligence in managing AI tools. The first UKAS-accredited certification body, BSI, began offering certification in January 2026. Employers who implement this standard may find it easier to meet Local Law 144 requirements. The standard helps ensure that AI tools are reviewed and audited consistently, reducing the risk of unintentional bias. The certification process involves an assessment of the organisation’s AI governance framework, including policies, procedures, and audit practices.