Racist Computers? EEOC issues guidance for companies using artificial intelligence for employee recruitment.
Posted on Jun 15, 2023 on Employment, Employment-at-Will / Restrictive Covenants, Legal Updates, Legislative Updates, Musings from an HR Attorney, News by
Are you using artificial intelligence (“AI”) to support HR-related activities? If so, have you asked your AI vendor how they protect against bias? If not, you should. The Equal Employment Opportunity Commission (“EEOC”), in a recent announcement, made one thing very clear: employers will be liable for any discrimination produced by AI.
A 2022 study by the Society for Human Resource Management (“SHRM”) shows 1 in 4 organizations report using artificial intelligence (AI) to support HR-related activities, including recruitment and hiring. Of the organizations who use these tools, only 2 in 5 say their AI vendors are transparent about the steps taken to ensure the tools protect against bias.
SHRM’s study suggests HR managers who utilize AI hiring tools do not understand the risks. Additionally, AI vendors will not usually disclose their testing methods and will require companies to sign indemnification clauses, thus leaving employers on the hook for potentially expensive EEOC litigation.
How can you prevent this?
If your organization is using AI, ask the AI vendor what steps have been taken to evaluate whether the AI causes a lower selection rate for individuals protected by Title VII. If the AI does cause a lower selection rate, the organization should explore whether the AI meets the EEOC’s four-fifth rule. The four-fifths rule states that a protected class may be adversely impacted if the selection rate of the protected class is less than four-fifths (80 percent) of the selection rate of the non-protected class. However, this is a rule of thumb and does not always insulate employers from liability.
Finally, employers may want to consider implementing a regular AI oversight process. AI oversight should routinely test for potential adverse impacts. If there is any impact, the employer should consider whether the impact is job-related and consistent with business necessity. You should also consider whether there are any alternatives. If there is no business necessity (or if there are reasonable alternatives that avoid such a result), employers should require their AI vendor to address and correct these issues.
Brody and Associates regularly advises management on complying with the latest local, state and federal employment laws. If we can be of assistance in this area, please contact us at info@brodyandassociates.com or 203.454.0560