NYC Local Law 144 requires employers to publish a public summary of their automated employment decision tool (AEDT) bias audit results. This publication requirement is separate from the independent audit itself and from the candidate notice requirement. Many employers focus on completing the audit but overlook the specific content and formatting requirements for the public summary. This article provides a practical guide to what the summary must contain, where to publish it, and how to maintain it over time.
What This Means
The public summary requirement under LL144 is designed to create transparency around how automated tools affect employment decisions. The law assumes that candidates and the public have a right to know whether an AEDT produces disparate outcomes across protected categories. By requiring public disclosure, the law creates accountability and enables candidates to make informed decisions about whether to request an alternative selection process. The summary must be published before the employer uses the AEDT. This means you cannot conduct an audit, begin using the tool, and post the summary later. The publication must precede the use. If you are already using an AEDT and have not yet published a summary, you are technically in violation of the law even if you have completed the independent audit. The location of the summary matters. It must be published in the careers or jobs section of the employer's website. If your company does not have a dedicated careers page, the summary must be placed in a similarly prominent location where candidates are likely to find it. Hiding the summary in a privacy policy, terms of service, or footer link is insufficient. The DCWP has not issued detailed guidance on what constitutes an adequate location, but the spirit of the law is clear: the summary must be easily discoverable by candidates. For employers with multiple AEDTs, the law requires a separate summary for each tool. You cannot combine results from different tools into a single summary unless the tools were evaluated together in a single audit. If you have separate audits for separate tools, you must publish separate summaries. This is particularly relevant for large employers that may use different tools for different stages of the hiring process. The public summary requirement is purely a transparency obligation. It does not require the employer to change the tool, stop using the tool, or take any specific remedial action based on the audit results. However, the fact that the results are public creates reputational and legal incentives to address any disparities identified. Candidates, journalists, and regulators can all access the summary, and unfavorable results may attract scrutiny.
Key Requirements
Date of the Audit.
The summary must include the date on which the most recent bias audit was completed. This date is critical because the audit must be no more than one year old. If a candidate sees a summary dated more than a year ago, they can infer that the employer may not be in compliance. Employers should update the date whenever a new audit is completed and remove old summaries that have expired.
Distribution of Categories.
The summary must list the demographic categories that were considered in the audit. These categories are race, ethnicity, and sex, including intersectional combinations. The DCWP rules require that the summary disclose the distribution of these categories, meaning the employer must show which categories were tested and how many individuals fell into each category. This does not mean publishing individual names or raw data, but it does mean showing the breakdown of the candidate or employee pool by category.
Selection Rates.
The summary must include the selection rate for each category. The selection rate is the percentage of individuals in a given category who were selected or advanced by the AEDT. For example, if 100 Asian women applied and 20 were selected, the selection rate for Asian women is 20%. These rates must be disclosed for every category tested, not just aggregate rates.
Impact Ratios.
The summary must include the impact ratio for each category compared to the most selected category. The impact ratio is calculated by dividing the selection rate of a given category by the selection rate of the most selected category. For example, if the most selected category has a 30% selection rate and another category has a 15% selection rate, the impact ratio for the second category is 0.50 or 50%. The summary must include these ratios for all categories.
Accessibility.
The summary must be accessible to the public without requiring a login, application, or payment. It must be published in plain text or HTML, not buried in a PDF or video. The DCWP has not specified accessibility standards for the summary itself, but best practice suggests using clear formatting, readable fonts, and a structure that makes the categories, rates, and ratios easy to understand.
Annual Update.
The summary must be updated whenever a new audit is conducted. If the employer stops using an AEDT, the summary should be removed or updated to indicate that the tool is no longer in use. There is no specific requirement to keep historical summaries, but some employers choose to archive them for internal compliance records.
Practical Steps
Format the Summary Clearly.
Use a simple table or structured list to present the audit date, categories, selection rates, and impact ratios. Avoid dense paragraphs or legal jargon that obscures the data. The goal is transparency, not obfuscation. A well-formatted summary should allow a candidate to understand the audit results in under a minute.
Include All Required Categories.
Do not omit intersectional categories or categories with small sample sizes. The DCWP rules require testing across all applicable categories. Even if a category has a small sample size, it should be included in the summary with a note if the sample size is insufficient for reliable statistical analysis. Omitting categories because the results are unfavorable is a misrepresentation and may violate the law.
Publish in the Right Location.
Create a dedicated page or section within your careers site for the LL144 summary. Link to it from your main careers page, job postings, and any AEDT notice pages. The link should be labeled clearly, such as "AEDT Bias Audit Summary" or "Local Law 144 Compliance." Do not hide it behind generic labels like "Legal" or "Policies."
Coordinate with Your Auditor.
Ask your independent auditor to provide the summary data in a format that is ready for publication. Some auditors will prepare a public-facing summary as part of their deliverable. Others will provide only the raw data and expect the employer to format it. Clarify this expectation before the audit begins to avoid delays.
Review Before Publishing.
Have your legal counsel and compliance team review the summary before it goes live. Check that all required elements are present, that the math is correct, and that the formatting is clear. A simple error in a selection rate or impact ratio can create confusion and compliance risk. A brief review process prevents these issues.
Monitor for Changes.
If your AEDT changes, if you add a new tool, or if your candidate pool shifts significantly, you may need an updated audit and a new summary. Build a process for reviewing your AEDT inventory and summary status at least quarterly. This ensures that you do not accidentally use an expired audit or fail to publish a summary for a new tool.
Related Resources
HR AI Compliance
Frequently Asked Questions
What must an LL144 public summary include?
An LL144 public summary must include the date of the most recent bias audit, the distribution of demographic categories considered (race, ethnicity, and sex, including intersectional categories), the selection rates for each category, and the impact ratios for each category compared to the most selected category.
Where must the public summary be published?
The public summary must be published in the careers or jobs section of the employer's website. It must be easily accessible to the public before the employer uses the AEDT. The link must be clear and not hidden behind login screens or application portals.
How often must the public summary be updated?
The public summary must be updated whenever a new bias audit is conducted, which must occur at least annually. The audit must be completed no more than one year before the AEDT is used. If the audit expires, the employer must remove or update the summary to reflect the current audit status.
Can an employer publish only favorable results?
No. The public summary must accurately reflect the full audit results. Selectively publishing only favorable categories or omitting unfavorable impact ratios would likely constitute a misrepresentation. The DCWP rules require disclosure of the distribution of categories and the corresponding selection rates and impact ratios.
What if the audit finds no disparate impact?
Even if the audit finds no statistically significant disparate impact, the employer must still publish the summary with the actual selection rates and impact ratios. A finding of no disparate impact does not eliminate the publication requirement. The summary must still include the date, categories, and metrics. Review Your LL144 Public Summary Our free AI Regulatory Readiness Assessment includes a public summary checklist as part of its 43-control evaluation.