As artificial intelligence (“AI”) tools like ChatGPT, Bing and Baird become more popular, legislation addressing such technology is inevitable. New York City has implemented the first law in the United States that regulates AI and automated decision-making tools in the employment space.
Legislative History and Proposed Rules
On November 10, 2021, the New York City council passed the New York City Local Law 144, which regulates AI and automated decision-making tools for employee hiring and promotions. The bill was signed into law on December 11, 2021. Originally, this law was supposed to go into effect on January 1, 2023, but the New York City Department of Consumer and Worker Protection (DCWP), which enforces this law, received a high volume of public comments on the proposed implementing rules from academics, employers and civil society organizations, which delayed the effective date of the law. The DCWP published proposed rules in September 2022 and December 2022, which received further feedback. Finally, on April 6, 2023, the DCWP issued the final rule, effective July 5, 2023.
DCWP’s Final Rule
The Final Rule builds on the previous rules. Under the final rule, the use of Automated Employment Decision Tools (AEDTs) is prohibited for employment decisions unless specific conditions are met. An AEDT is defined as a computer-based tool that uses machine learning, statistical modeling, data analytics or AI to substantially help with employment decisions.
As a prerequisite to an employer’s use of an AEDT, the computer tool must have undergone a bias audit within one year of its use. Information about the bias audit must be publicly available, and certain notices must be provided to employees or job candidates. Even tools that do not use AI but rely on less sophisticated statistical modeling or data analytics can fall under this law. Such tools must undergo an impartial evaluation by an independent auditor to satisfy the bias audit. The bias audits must calculate the selection rate for each race/ethnicity and sex category required by the U.S. Equal Employment Opportunity Commission (EEOC) and compare it to the impact ratio using a specific formula.
Before using an AEDT, an employer must publicly disclose on the employment section of its website, in a “clear and conspicuous manner,” the date of the most recent bias audit for the AEDT in use and a summary of the results. It must also provide the distribution date of the AEDT. This information must be posted for at least six months after the latest use of the AEDT.
Additionally, employers must inform individuals about how they can request an alternative selection process or a reasonable accommodation under other laws, if available. The DCWP has outlined various methods for providing notice, such as through job postings, the company’s website, or via U.S. mail or email. Failure to comply with any of the above can lead to fines starting at $500 with a maximum penalty of $1,500.
AI’s Growing Trend
This new AI law is part of a growing trend of AI bias laws seen in states like California, Illinois, New Jersey and Maryland. On the federal level, the EEOC launched its Artificial Intelligence and Algorithmic Fairness Initiative and recently provided guidance on the use of AI in employment selection procedures. New York City’s AI law requires employers to determine whether they have an AEDT in place, conduct appropriate bias audits, and work with employment counsel to ensure compliance with each step.