AI technology is advancing quickly, and regulators are trying to keep up with it, which often leaves businesses caught between a rock and a hard place. If they withhold the use of AI, they risk falling behind competitors, while using AI means they must comply and stay up to day with all the regulations.
However, it is hard to deny that despite all the advantages AI can bring to businesses, it also comes with its risks, such as bias in AI. That is why certain regulations are necessary.
Addressing these concerns, New York City has implemented Local Law 144, effective July 5th, which mandates auditing AEDTs for biases. In this blog post, we will explore the implications of this law and its potential impact on businesses.
What are AEDTs?
Under Local Law 144, AEDTs encompass tools that substantially “assist or replace an employer’s discretion in making employment-related decisions”. This definition includes tools that:
- score, classify, rank, or evaluate job applicants or employees based on specific factors;
- use multiple factors in the decision-making process but place a greater weight on one particular factor over the others;
- use multiple factors in the decision-making process, with one particular factor being able to overrule the conclusions of all other factors.
Employers are encouraged to engage an independent auditor to conduct a bias audit before using any software for employment decisions to ensure fair practices.
What does it mean for businesses?
Local Law 144 introduces crucial changes for employers in New York City. Employers and agencies must audit any automated employment tools annually and publicly disclose data and summary from the bias audit.
It is worth noting that bias audits are necessary not only for AEDTs used in final hiring decisions but also for tools employed during early screening stages. Employers will be obligated to disclose details in the summary of bias audit results, including
- “the source of explanation” of the data used to conduct the bias audit,
- the date of the last audit,
- the number of applicants and candidates,
- the selection of scoring rates,
- the impact ratio by race/ethnicity, sex, and intersectional categories,
- the number of individuals with unknown race/ethnicity and sex that were assessed by the tool.
Furthermore, employees and candidates must be notified about the use of AEDTs in the screening process at least ten business days before implementing the tool. This is a significant departure from previous practices requiring greater transparency from businesses. Job applicants will now have the opportunity to gain insights into the decision-making process.
Failure to comply with these regulations, including the notice requirements, can result in fines ranging from $500 to $1,500 for each violation.
Can we expect similar laws to follow?
The United States Equal Employment Opportunity Commission (EEOC) has released new rules to prevent discrimination by employers when using AI tools during the hiring process. These rules are expected to apply nationwide. Similar laws to the NYC Local Law 144 are expected to follow in California and New Jersey.
Additionally, the EU AI Act covers a wide range of business operations and includes regulations concerning AI tools used for employment decisions. High-risk AI applications, including AI employment decision tools, will be subject to auditing requirements once the EU AI Act is enforced.
Modern-day businesses heavily rely on automated tools for candidate screening, as they increase efficiency and expedite recruitment. However, it is crucial to ensure that these tools do not disadvantage job candidates or employees due to biases embedded within AI algorithms. New York City’s Local Law 144 aims to benefit job seekers and employees by promoting transparency and fairness. By adhering to such regulations, businesses can demonstrate their commitment to fairness and integrity in their employment practices. Ultimately, the law enables businesses to tap into a wider pool of candidates while ensuring equal opportunities and fostering a more inclusive workforce.