EEOC Goes After An AI As An Employer

There's an interesting employment law case developing in California (it's always California) involving the status of a software company as an employer or employment agency for one of its clients.  The case potentially represents a major expansion of employment law liability for companies supplying services that involve employee screening, hiring, or any other kind of concrete employment decision. The implications for artificial intelligence programs in particular look to be very serious. 

The facts are straightforward. An individual made over 100 applications to various employers that used an online application template provided by a company named Workday.  Potential hires would apply online to companies through various hiring sources (e.g., LinkedIn) and were then directed to the Workday app where they would input a resume and personal details, and in some instances take an aptitude test. The plaintiff in this case was rejected for all of his applications (we have no indication of whether he was qualified for the jobs for which he applied). He is now claiming that these rejections were based on his race and disability, which he apparently was revealing as a matter of course in the application process. 

Important safety tip here- while I have not seen the Workday template, I'm reasonably certain that it does not ask for the race and a list of disabling conditions on the part of applicants, At least as part of an initial hiring process. How and why this information was being input at this stage raises an interesting question about whether these were bona fide employment applications. 

In any event, the Workday app reviews the templated information and then forwards qualifying applications on to the employer. There is no interview interaction or any other kind of personal evaluation of the applicant- the software simply screens the application for job requirements and determines if this is a potentially good match based on the criteria from the employer. 

Given its relatively mundane function, imagine Workday's surprise when it found itself in the crosshairs of an employment discrimination lawsuit filed not by one of its employees, but by one of its customer's employees. And to make matters worse, the EEOC has now jumped into the fray, arguing that the company is in fact acting as an employment agency and therefore is subject to Title VII for the employment decisions that are, in essence, being made as a result of the criteria Workday receives from its customers. 

The motion detailing these arguments is here. The implications are much broader than simply whether Workday is acting as an employment agency. Many companies are moving to artificial intelligence algorithms to assist them in making employment decisions involving hiring, firing, promotions, etc. Under the EEOC's theory, the manufacturers of these software systems could find themselves in the Commission's sights either as employers outright or as some kind of quasi-employer. I'm reasonably certain the Commission is attacking AI systems as discriminatory across the board because their widespread use will likely mean the end of affirmative action as we know it.

Imagine you manufacture an AI system that helps employers assess talent for internal promotions. Suddenly you are subject to Title VII liability not only for the employees in your own company, but for the tens of thousands of employment decisions for which your application is used throughout the United States. Sound crazy? I agree. But this is the logical outgrowth of the EEOC's position in this Workday case. 

Stand by for more exciting news from Northern California.