On July 5, 2023, New York City is poised to begin enforcement of a law initially passed in 2021 that requires employers to take certain steps before implementing an automated employment decision tool (AEDT). The law reflects a growing focus on the use of artificial intelligence (AI) in employment decisions and addresses concerns pertaining to the use of AI, including that biased algorithms could create a disparate impact on certain candidates or employees, even without an employer’s intent to discriminate.
NYC 144 Requirements
As previously detailed here, NYC Local Law 144 of 2021 (NYC 144) prohibits employers and employment agencies from using an AEDT to make a hiring or promotion decision unless they conduct a bias audit on the tool annually, publish an audit summary, and provide certain notice to candidates and employees about the tool. New York City previously deferred enforcement of the law pending final regulations, which were published on April 5, 2023, and the law will become effective July 5, 2023. These regulations provide further insight into what the law covers and how an employer should conduct a bias audit.
The New York City law applies to an employer’s or employment agency’s use of an AEDT to screen candidates for hiring or employees for promotion. The law does not cover other employment decisions, such as compensation, retention, and termination decisions. In the context of hiring, the law applies with respect to tools used to screen candidates who have actually applied for a specific job and not to tools used to identify potential candidates who have not yet applied for a position. Without further guidance provided by the city, the prudent course is to comply with the law with respect to those candidates and employees who live or will work in New York City.
An AEDT means any computational process, derived from machine learning, statistical modeling, data analytics, or AI, that issues simplified output, including a score, classification, or recommendation, that is used to substantially assist or replace discretionary decision making for employment decisions. An AEDT does not include junk email filters, firewalls, antivirus software, calculators, spreadsheets, databases, or data sets.
The law defines a “bias audit” as “an impartial evaluation by an independent auditor.” Specifically, an independent auditor must be an objective individual or group that has not been involved in the AEDT’s use, development, or distribution and is not employed by or financially interested in the employer using the AEDT or the vendor that develops or distributes the AEDT. Accordingly, the AI software vendor does not qualify as “independent” under the law and could not conduct the bias audit itself. An AEDT also cannot be used if more than one year has passed since the most recent bias audit of the tool.
The regulations further detail the content of what the bias audit must include, with specifications regarding the technical processes and calculations to be conducted. The bias audit must include calculations of the selection (or scoring) rate for each category, including sex, race/ethnicity, and intersectional categories, and the “impact ratio” of each category. “Impact ratio” is defined as either the selection rate for a category divided by the selection rate of the most-selected category or the scoring rate for a category divided by the scoring rate for the highest-scoring category. The regulations also specify that historical data–the data collected during an employer’s prior use of an AEDT–must be used for the bias audit unless the employer has never used the AEDT or there is insufficient historical data available to conduct a statistically significant audit.
Before the use of an AEDT, an employer must publicly post on its website certain information, including the date of the most recent bias audit, a summary of the results, and the date it began using the AEDT. An employer must also notify candidates and employees, via certain acceptable methods, no less than 10 business days before use of the AEDT (1) that an AEDT will be used; (2) of instructions for how an individual can request an alternative selection process or a reasonable accommodation under other laws, if available; and (2) of the job qualifications and characteristics the AEDT will use in its assessment. The regulations provide that nothing in them requires an employer to provide an alternative selection process. The employer must additionally provide information about the type of data collected for the AEDT, the source of the data, and the employer’s data retention policy and post instructions for how to make a written request for the information.
The New York City Department of Consumer and Worker Protection will hold an “Educational Roundtable with Business Advocates/Employers” on May 22, 2023, to provide an overview of the regulations and field questions. It anticipates issuing a FAQ document for future education.
Penalties for Violations
Any person who violates the law is liable for a civil penalty of (1) not more than $500 for a first violation and each additional violation occurring on the same day as the first violation and (2) not less than $500 nor more than $1,500 for each subsequent violation. Each day on which an AEDT is used in violation of the law constitutes a separate violation, and the failure to provide any of the required notices to candidates or employees constitutes a separate violation. A private right of action is available under the law.
In line with the growing interest in AI and employment, the U.S. Equal Employment Opportunity Commission announced an initiative last year to ensure that AI used in hiring and other employment decisions complies with federal anti-discrimination laws.
On January 1, 2020, the Artificial Intelligence Video Interview Act took effect in Illinois, requiring employers conducting AI analysis of applicant-submitted videos to comply with certain notice and other requirements. Effective October 1, 2020, Maryland enacted a law pertaining to employers’ use of facial recognition technology during preemployment job interviews. Legislation is currently proposed in a number of cities and states concerning the use of AI in the employment context, including in California and Washington, D.C.
New York City employers that use or intend to use AI-driven tools in hiring and promotional processes should take steps now to facilitate compliance with NYC 144 in advance of the July 5, 2023, enforcement date.
If you need assistance navigating the complex employment laws governing the workplace, please contact Lowenstein Sandler’s Employment Counseling & Litigation practice group.