On Friday, September 23, 2022, the New York City Department of Consumer and Worker Protection (“DCWP”) released a Notice of Public Hearing and Opportunity to Comment on Proposed Rules related to its Automated Employment Decision Tool law (the “AEDT Law”), which goes into effect on January 1, 2023. As we previously wrote, the City passed the AEDT Law to regulate employers’ use of automated employment decision tools, with the aim of curbing bias in hiring and promotions; as written, however, it contains many ambiguities, which has left covered employers with open questions about compliance.
The proposed rules are intended to clarify the requirements for the use of automated employment decision tools within New York City, the definitions of key terms in the AEDT law, the notices to employees and applicants regarding the use of the tool, the bias audit for the tool, and the required published results of the bias audit.
The DCWP’s public hearing on the proposed rules and deadline for comments are October 24, 2022. Although the proposed rules may be modified prior to adoption, the following summarizes the key provisions.
“Substantially assist or replace discretionary decision making”
The AEDT Law applies to an automated decision tool that is used “to substantially assist or replace discretionary decision making.” It does not, however, specify the type of activities that constitute such conduct or what particular AI-powered employment tools are covered by the law.
The proposed rules attempt to provide guidance on this issue by defining “substantially assist or replace discretionary decision-making” as one of the following actions:
- relying solely on a simplified output (score, tag, classification, ranking, etc.), without considering other factors; or
- using a simplified output as one of a set of criteria where the output is weighted more than any other criterion in the set; or
- using a simplified output to overrule or modify conclusions derived from other factors including human decision-making.
“Bias Audit”
Pursuant to the AEDT Law, before using an automated employment decision tool, a covered employer or employment agency must subject the tool to a “bias audit” no more than one year prior to the use of the of the tool. The law explains that “bias audit” means an “impartial evaluation by an independent auditor,” but does not otherwise specify who or what constitutes an “independent auditor” or what the “bias audit” must contain. The proposed rules address these gaps.
First, the proposed rules define “independent auditor” as “a person or group that is not involved in using or developing an [automated employment decision tool] that is responsible for conducting a bias audit of such [tool].” This definition does not specify that the auditor must be a separate legal entity from the creator or vendor of the tool and therefore suggests that it may be acceptable for the auditor to be employed by the organization using the tool, provided the auditor does not use and has not been involved in developing the tool.
Second, the proposed rules state that the required contents of a “bias audit” will depend on how the employer or employment agency uses the tool.
If the tool selects individuals to move forward in the hiring process or classifies individuals into groups, the “bias audit,” at a minimum, would need to:
- calculate the selection rate for each category;
- calculate the impact ratio for each category; and
- where the tool classifies candidates into groups, the bias audit must calculate the selection rate and impact ratio for each classification.
If the automated employment decision tool merely scores candidates, the “bias audit” at a minimum, would need to:
- calculate the average score for individuals in each category; and
- calculate the impact ratio for each category.
The preamble to the proposed rules makes clear that DCWP intends these calculations to be consistent with the Uniform Guidelines on Employee Selection Procedures (“UGESP”), 29 C.F.R. § 1607.4, and borrows concepts from the framework established by the UGESP in the definitions of “impact ratio” and “selection rate.”
Under the AEDT Law, upon completion of a bias audit, and prior to using the automated employment decision tool, covered employers and employment agencies must make the date and summary of the results of the bias audit publicly available on the careers or job section of their website in a clear and conspicuous manner. The proposed rules clarify that publication may be made via an active hyperlink to a website containing the required information, as long as the link is clearly identified as linking to the results of the bias audit. The required information must remain posted for at least six months after the covered employer or employment agency uses the tool for an employment decision.
Required Notices
The AEDT Law also specifies that employers and employment agencies must notify candidates for employment and employees who reside in New York City as follows:
- at least ten business days prior to using an automated decision tool, that such a tool will be used to assess or evaluate the candidate or employee, and allow the individual to request an alternative selection process or accommodation;
- at least ten business days prior to use, the job qualifications and characteristics that the tool will use in the assessment or evaluation; and
- if not disclosed on the employer or employment agency’s website, information about the type of data collected for the tool, the source of such data, and the employer or employment agency’s data retention policy shall be available upon written request by the individual and be provided within thirty days of the written request.
Covered employers and employment agencies have expressed concern about the practical and administrative difficulties of providing the above notices in the fast-paced environment of today’s recruiting and hiring.
In apparent response to these concerns, the proposed rules clarify that the employer or employment agency may provide the notices required by paragraphs (1) and (2) by:
- (a) in the case of candidates, including notice on the careers or jobs section of its website at least ten business days prior to the use of the tool, and (b) in the case of employees, including notice in a written policy or procedure that is provided to employees at least ten business days prior to use;
- including notice in a job posting at least ten days prior to using the tool; or
- (a) in the case of candidates, providing notice via U.S. mail or email at least ten business days prior to use of the tool; and (b) in the case of employees, providing written notice in person, via U.S. mail, or email at least ten business days prior to use.
In short, under the proposed rule, an employer or employment agency could comply with the AEDT Law by providing the required notice when first posting the job.
With respect to the notice requirement in paragraph (3), the proposed rules state that an employer or employment agency must provide notice to covered individuals by including notice on the careers or jobs section of its website, or by providing written notice in person, via U.S. mail, or by email within 30 days of receipt of a written request for such information. If notice is not posted on the website, the employer or agency must post instructions for how to make a written request for such information on its careers or job section of the website.
Finally, although the AEDT Law requires an employer or employment agency to allow covered individuals to request an alternative selection process, the proposed rules state that nothing requires an employer or employment agency to provide an alternative selection process.
Conclusion
Epstein Becker & Green, P.C. will continue to monitor developments regarding the AEDT Law and will post an update when the DCWP issues final rules. In the meantime, please feel free to contact the authors if you have any questions regarding complying with the AEDT Law, would like to comment on the proposed rules, or would like to inquire about the use and regulation of workplace artificial intelligence tools in general.
Blog Editors
Authors
- Member of the Firm
- Member of the Firm