Sex, Money, and Machines: How the Fate of AI Pay Equity Software Depends on the Resolution of a Circuit Split

Katherine Kohan

ABSTRACT

Since the Equal Pay Act (EPA) was enacted, employers faced with a pay discrimination claim have attempted to use an employee’s prior compensation to justify a sex-based pay disparity. Under the Equal Pay Act, however, employers can only escape liability where the disparity can be attributed to one of four exceptions. Employers have argued that prior compensation falls under the fourth exception as a “factor other than sex.” The federal courts of appeals have taken different approaches when deciding whether an employee’s prior compensation is a permissible “factor other than sex” that employers may use to defend against a claim arising under the Equal Pay Act. Amid this developing landscape, state and local legislatures are rolling out pay equity legislation of their own. Many employers have turned to artificial intelligence algorithms to help them comply with multiple levels of pay transparency law in their jurisdiction. These algorithms often contain features that make salary recommendations based on compensation data from the labor market as a whole, and from within the employer’s organization. Because AI models may be trained by prior compensation data, legislative or judicial resolution of the circuit split could impact an employer’s ability to confidently rely on pay equity algorithms. This Note analyzes employer use of pay equity algorithms in light of the developing law on the issue and argues that the circuit split should be resolved to allow good faith employer reliance on AI pay equity software.

---