The U.S. Equal Employment Opportunity Commission (EEOC) has issued new guidance on using artificial intelligence (AI) in employment practices under Title VII of the Civil Rights Act of 1964. Specifically, the EEOC advises that employers may not rely on representations by vendors that their AI tool complies with Title VII. Instead, they must ensure that the tool has no adverse discriminatory impact. Otherwise, the employer could be held liable for discrimination.
Hiring Selection Tools and Procedures Must Be Neutral, Job-Related, and Consistent with Business Necessity – Including AI
Neutral tests or selection procedures used by employers during hiring processes, including AI tools, that have a disparate impact based on race, color, religion, sex, or national origin, must be related and necessary to the job or business, or Title VII prohibits their usage. Furthermore, even if the procedures are job-related and consistent with business necessity, employers must consider all available alternatives to avoid creating an illegal disparate impact.
All too often, employers are unaware of the applicability of Title VII to AI tools and the risks of using them. AI vendors also routinely refuse to disclose their testing methodology and disclaim all liability for potential adverse impacts in their licensing contracts. As a result, employers may want to inquire specifically whether the vendor has evaluated the tool for Title VII compliance or whether its use has resulted in any substantially lower selection rates for individuals with characteristics protected by Title VII. Although it is the vendor that has developed the AI tool, the EEOC guidance makes it clear that the employers carry the burden of legal compliance with Title VII.
AI Tools Should Comply with the Four-Fifths Rule
The EEOC guidance explains that the four-fifths rule for applicant selection rate applies to AI tools, just as it applies to other employer decision-making procedures. Therefore, employers may wish to ask vendors if the AI tool has been tested using the four-fifths rule.
Under the four-fifths rule, which equates to 80%, employers can measure whether a selection procedure or decision-making tool has an adverse impact on a specific group of individuals with a protected characteristic. If the use of the procedure or tool causes a selection rate for individuals in the group that is substantially less than the selection rate for individuals in another group, and the ratio of those two selection rates is lower than four-fifths or 80%, then it could be evidence of discrimination. Still, the EEOC points out that the four-fifths rule is not foolproof and may not be an appropriate measure of disparate impact in every case.
HR Oversight of AI Tools
Since the employer, not the vendor, will be liable for any Title VII violations resulting from using AI tools in employment hiring practices, employers must establish adequate oversight of those tools. Specifically, a chief AI officer should oversee all such tools and routinely evaluate them for potentially adverse disparate impact.
HBL has experience in all areas of benefits and employment law, offering a comprehensive solution to all your business benefits and HR/employment needs. We help ensure you are in compliance with the complex requirements of ERISA and the IRS code, as well as those laws that impact you and your employees. Together, we reduce your exposure to potential legal or financial penalties. Learn more by calling 470-571-1007.
Hall Benefits Law, LLC
Latest posts by Hall Benefits Law, LLC (see all)
- Aetna and MLB Face ERISA Suit Over Denied Mental Health Claims - September 27, 2023
