Adam Forman, Nathaniel Glasser Quoted in “NYC AI Bias Bill May Be Compliance Headache for Employers”

A nearly enacted New York City law would require businesses to conduct bias audits before using automated hiring tools and notify city-based applicants about how the technology works. The proposed legislation raises tricky compliance questions, and foreshadows what similar laws elsewhere might look like, attorneys say.

The bill, known officially as Int. 1894-2020, was passed by the New York City Council on Nov. 10 by a 38-4 vote and is currently awaiting Mayor Bill de Blasio's signature. If finalized without major hiccups, the law will take effect in January 2023.

Although businesses have time before the city's potential new rules are scheduled to kick in, management-side attorneys say it is the latest example of lawmakers increasingly wading into employers' use of artificial intelligence-based tools — legislative initiatives that will only gain steam as the technology becomes more prevalent. …

How Tricky Will It Be to Sort Out Residency?

Besides bias audits, the city's bill includes various other elements, including that businesses notify any applicant "who resides in the city" both that an automated tool will be used and what criteria it will take into account to weed out applications for promotions or jobs. Those notices must be provided at least 10 days before the tool is used, and job candidates will have a chance to ask for an "alternative selection process" or accommodation. Companies found to be in violation of the law would be subjected to monetary civil penalties.

Those requirements, however, could potentially raise numerous issues for employers, one of which is figuring out exactly who should receive a notice. 

"There are definitely going to be compliance questions as employers in and around New York City use automated employment decision tools," said Nathaniel Glasser, co-leader of Epstein Becker Green's artificial intelligence practice group.

For example, Glasser questioned whether the 10-day notice requirement applies to "each and every use" of a particular tool and how employers will be able to meet that obligation "when you have a fast-moving interview and hiring process."

Additionally, he questioned how the law will apply if, for example, a New York City-based company is hiring for a position in New Jersey where the applicant pool likely includes a mix of city residents and nonresidents.

"It only applies to candidates that reside in the city, which will actually be a complicating factor for employers that are hiring into New York City roles because employers may not know as an initial matter where somebody resides," Glasser said. "So, employers are going to have to make a decision as to whether or not they obtain a candidate's residency before giving the notice. You can give the notice only to those candidates that reside in the city or end up giving the notice to everybody for a particular role regardless of where they might live."

However, Epstein Becker member Adam Forman, who also co-chairs the firm's AI practice, cautions that in making those considerations, employers should keep in mind that asking people where they live could itself result in unlawful disparate impact.

"It puts the employer in a pickle," Forman said. "As a practical proposition, if you're an employer recruiting citizens of New York City and seeking to use [a] tool, even though you'll get people from outside the city, it may be an all or nothing proposition. And when I say that, I really mean an all proposition — you want to comply with local law [and] it just means you may be giving information to those who otherwise aren't legally required to receive it."

How Many Other Jurisdictions Will Jump Aboard?

While New York City's proposed law is itself notable, attorneys said it isn't an outlier when it comes to lawmakers increasingly eyeing AI as an area that warrants more oversight. …

Besides drawing attention from states and municipalities, employers' use of AI is increasingly catching the attention of federal regulators.

The U.S. Equal Employment Opportunity Commission recently announced a new initiative aimed at ensuring that AI and algorithms that are used in hiring and employment decisions don't run afoul of federal civil rights laws.

As part of its program, the EEOC said it would create an internal working group to coordinate the agency's activities, host listening sessions and issue technical guidance for using AI to make employment decisions.

"I do think that [New York City's law] portends future actions by other jurisdictions, whether it will be on a local level, such as a municipality, city, township, state level ... or a federal level," said Forman of Epstein Becker. "With the EEOC announcing a task force recently to take a closer look, I think we will see regulation of an employer's use of artificial intelligence or as the New York City bill calls it, 'automated tools.'"