Investigating New York City’s Automated Employment Decision Tools Law

Although recent public attention has focused on generative artificial intelligence (AI), its use in hiring and promotion screening is already widespread in employment. With job postings and application processes increasingly happening online, it helps human resources professionals make sense of the data. According to a study conducted by the Society for Human Resource Management (SHRM),[1] Almost a quarter of organizations use automation or AI to support HR-related activities such as recruiting, hiring, and promotion decisions, and that number is likely to grow. The same study found that one in four organizations plans to start or increase the use of automation or AI in recruiting or recruiting over the next five years, and one in five I know you’re planning to start using it or increase your usage. , automation or AI in performance management in the next five years.

Using AI tools for HR purposes can influence workers at every stage of employment, from applicant screening and performance counseling to promotion and disciplinary termination decisions. Different approaches to regulating these tools are emerging. We discussed the proposed Anti-Robot Bosses Act, which would ban the use of “automated decision-making systems” in some cases and give employees certain opt-out rights. Similarly, we reported on the EEOC’s recent guidance regarding the use of automated systems, focusing on how the use of these tools may violate Title VII.

New York City’s approach to regulating these tools aims to make their use more transparent. Local Law 144 of 2021 (“Local Law 144”) went into effect on January 1, 2023 and is one of the few laws in recent years to have passed the legislative process regulating AI in the employment context.[2] We have previously written about Local Law 144, from its passage to its implementing regulations (the “Final Rules”) and the New York City Department of Consumer and Worker Protection’s (“DCWP”) release of FAQs regarding the law (“). I have written. FAQ”). This law has been in effect for several months, starting July 5, 2023, so now is a good time to consider its requirements.

In general, Local Law 144 requires covered employers and employment agencies that use “Automated Employment Determination Tools” (AEDTs) to meet two key requirements: (1) Confirm that AEDT is subject to a bias audit and publish the results. (2) Notify applicants and employees before using AEDT. The final rule and frequently asked questions will help employers and employment agencies comply with these requirements.

geographical scope of the law

The bias audit and notification requirements of Local Law 144 apply to the use of AEDTs “within the city.” DCWP’s FAQ can help further clarify this geographic restriction. First, the FAQ explains that compliance with bias audit requirements applies to: (1) Employers who use her AEDT for New York City-based jobs (including remote jobs associated with New York City addresses). (2) employment agencies located in New York City or recruiting for work in New York City; Regarding notice requirements, the FAQ clarifies that only applicants or employees who reside in New York City should receive the required notice.

What is an automated hiring decision tool?

AEDT is a computational process derived from “machine learning, statistical modeling, data analysis, or artificial intelligence” that produces “simplified output” (i.e. score, classification, or recommendation). “Hiring decisions” (i.e., screening candidates for employment or employees for promotion within the city). I wrote about the definition of AEDT and the final rule’s clarification of several terms within that definition. For example, the final rule clarifies that AEDTs are limited to those that (i) rely “only” on simplified outputs; (ii) Use the simplified output as a weighted factor over other criteria in the set. or (iii) use the simplified output to “overturn” conclusions drawn from other factors, including human decision-making. Some applications may include resume screening software or pre-employment assessment programs. Employee productivity and monitoring tools may also be covered to the extent that they generate scores or classifications used to screen employees for promotion.

Audit AEDT for potential bias

Within one year of the use of an AEDT, employers and employment agencies must ensure that a bias audit of the AEDT is conducted by an independent third party. The final rule clarifies the definition of “independent auditor” and requires that the auditor: Not be employed by a company that uses AEDT or a vendor that develops or distributes AEDT. We have no involvement in the use, development, or distribution of AEDT. have no direct financial interest or significant indirect financial interest in any employer or employment agency that uses the AEDT or in any vendor that developed or distributed the AEDT; Notably, however, the FAQ states that vendors can have independent auditors conduct bias audits of their tools. While this may reduce the burden on employers and recruitment agencies, the FAQ makes clear that the responsibility for ensuring a bias audit is completed remains with the employer or recruitment agency, not the vendor. .

Addressing public concerns that Local Law 144 is largely silent on what constitutes a “bias audit,” DCWP provided details in the final rule about what a compliant bias audit must include. . Specifically, the bias audit must include scoring and/or selection rates for each racial/ethnic and gender category reported in the EEO Component 1 report. You should also include an “impact ratio.” This means the selection/score rate for each category divided by the highest score/selected category. Influence rates indicate whether race/ethnicity and gender categories of applicants or employees are selected (or scored) at a higher rate than other categories.

A ‘summary’ of the bias audit results must be published on the employer or employment agency’s website. The summary must include the date of the AEDT’s most recent bias audit, the date of distribution of the AEDT, the source and description of the data used to conduct the bias audit, and the number of individuals rated “unknown” by the AEDT. It doesn’t have to be. ” category (i.e., people of unknown race/ethnicity and gender), number of applicants or candidates, selection rate or scoring rate, if applicable, and impact rate for all categories.

In its FAQ, DCWP confirmed that Local Law 144 “does not require any specific action based on the results of a bias audit.” However, the results of a bias audit may publicly reveal data that could support potential discrimination claims. As we recently reported, the U.S. Equal Employment Opportunity Commission (“EEOC”) alleges that companies had programmed their screen software to reject applicants over the age of 50, in violation of the Age Discrimination in Employment Act. I made peace with it.

Notification requirements

Local Law 144 requires 10 business days’ notice before using AEDT. Addressing the reality that many job applications are being made online, the final rule requires that at least 10 days prior to using the AEDT, the employer or employment agency must post a Provides that notice may be provided by posting a notice. Or email is sufficient. The FAQ also clarifies that notices do not have to be “position-appropriate.” Specifically for promotion candidates, the FAQ further clarifies that notification can be provided by a written policy or procedure. This may include, for example, your company’s employee handbook or privacy policy.

The notice itself must inform the employee or applicant that the AEDT is being used and the “job qualifications and characteristics” that the AEDT will analyze. You should also include instructions on how to request an alternative selection process or reasonable accommodation. Although the final rule clarifies that under local law 144 an employer or employment agency is not required to provide an alternative selection process in response to such a request, as we wrote here: Other laws, such as the Americans with Disabilities Act, may impose such requirements. .

Finally, local law 144 requires employers or employment agencies to provide information regarding AEDT-related data handling practices, specifically the types of data collected for AEDT, such You must provide the source of the data and the employer or employment agency information. Data Retention Policy. This information must be provided on the employer or recruitment agency’s website or “within 30 days of the written request,” and the final rule provides additional information on how to make such a written request. Instructions must be provided to the applicant or employee. However, if the employer or hiring agency explains to the candidate or employee that doing so would violate local, state, or federal law or obstruct a law enforcement investigation. You do not need to provide this information.

What penalties can an employer face for non-compliance?

Organizations that violate Local Law 144 will be liable for $500 per violation on the first day. Each violation after the first day will result in a civil fine of $500 to $1,500. Each day that an AEDT is used without complying with bias audit requirements is a separate violation, and so is failure to provide adequate notice to applicants or employees.

Overall picture

Regulation of AI in the employment context is rapidly evolving at the state and federal level. As noted above, federal efforts include the proposed Robot Boss Ban Act, EEOC guidance, and guidance from other regulatory agencies. Including FTC, CFPB, and Department of Justice. Additionally, state regulators such as the California Civil Rights Council and the California Privacy Protection Agency are working on regulations addressing the use of automated decision-making tools. New York State Senator Hoylman Segal also recently proposed expanded legislation to regulate the use of automated employment decision tools and electronic monitoring practices. Many of these proposals have similar overlapping requirements to Local Law 144. Therefore, understanding New York City law can help employers and employment agencies stay ahead of these ongoing regulatory efforts.

What should employers do now?

Employers or recruitment agencies with offices in New York City and companies that send employees to New York City should conduct an evaluation of the tools used for recruiting and promotion and the vendors that provide those tools. there is. Depending on the results of that assessment, employers and employment agencies may (1) have an independent auditor conduct a bias audit, (2) update the wording of the applicable job listing, or (3) update the employee handbook. You may need to revise your internal policies, including: and privacy policy. EBG’s dedicated AI team will assist you throughout this process.

*zoe reed A Summer 2023 Associate (not admitted to practice law) in the firm’s New York office contributed to the preparation of this post.

[1] SHRM, Automation and AI in HRavailable at

https://advocacy.shrm.org/SHRM-2022-Automation-AI-Research.pdf

[2] Illinois has also enacted legislation regarding the use of facial recognition and video interviewing technology used during the applicant interview process.