As we recently reported, on December 9, 2022, the New York City Department of Consumer and Worker Production (“DCWP”) announced that it was postponing enforcement of the Automated Employment Decision Tools (“AEDT”) law, until April 15, 2023, due to the high volume of public comments it received regarding its proposed rules.
Today, December 23, 2022, DCWP issued a notice with updated proposed rules to the AEDT law and has scheduled another public hearing on Monday, January 23, 2023 at 11:00 a.m. Eastern Standard time. DCWP is accepting comments on the updated proposed rules through January 23.
As summarized by DCWP, the proposed changes include:
- Modifying the definition of AEDT to ensure it is focused;
- Clarifying that an “independent auditor” may not be employed or have a financial interest in an employer or employment agency that seeks to use or continue to use an AEDT or in a vendor that developed or distributed the AEDT;
- Revising the required calculation to be performed where an AEDT scores candidates;
- Clarifying that the required “impact ratio” must be calculated separately to compare sex categories, race/ethnicity categories, and intersectional categories;
- Clarifying the types of data that may be used to conduct a bias audit;
- Clarifying that multiple employers using the same AEDT may rely upon the same bias audit so long as they provide historical data, if available, for the independent auditor to consider in such bias audit; and
- Clarifying that an AEDT may not be used if its most recent bias audit is more than one year old.
Whereas the definition of AEDT previously included tools that “use a simplified output to overrule or modify conclusions derived from other factors including human decision-making,” the updated rules no longer contain the word “modify.”
The updated proposed rules also contain several changes to the bias audit requirements. For example, if an AEDT scores candidates, the bias audit must calculate the median score for the full sample of applicants, the scoring rate for individuals in each protected category, and the impact ratio for each protected category. The updated rules no longer allow for the use of averages to compare protected categories. Notably, however, the updated rules have retained the requirement that these calculations be conducted based on sex categories (male vs. female), race/ethnicity categories, and other intersectional categories. If the requirement to compare intersectional categories remains in the final rules, employers and vendors may have to revisit their bias audit and data collection procedures to ensure compliance.
In addition, the proposal clarifies that a bias audit must rely on historical data of the AEDT, unless the auditor has insufficient historical data to conduct a statistically significant bias audit. If an auditor uses test data in lieu of historical data, the auditor must explain why it did not use historical data and describe how it generated and obtained the test data. Finally, an employer or employment agency must use its own historical data to conduct a bias audit, unless the employer or employment agency has never used the AEDT.
A person or group “capable of exercising objective and impartial judgment on all issues within the scope of a bias audit of an AEDT” must conduct the bias audit. The auditor cannot be involved in using, developing, or distributing the AEDT, or otherwise employed by or have a direct financial interest or material indirect financial interest in the user or vendor of the AEDT. In other words, the proposal eliminates the ability of employers that have their own auditing departments from conducting a bias audit for purpose of the AEDT law.