Artificial Intelligence can help write a resume, prepare someone for a job interview and search through thousands of open positions. It’s also being increasingly used by employers who manage job applications and reject candidates through human resources software.
But since many major software programs are equipped with the latest technology, AI frequently decides which job applicants an employer sees – and which ones never make it that far.
Now, a federal class action against Workday, one of the largest providers of human resources software, is putting a critical question before a court: If AI rejects someone for a job, who is legally responsible if it discriminates against them?
Whether it’s the company that made the hiring decision, the company that made the software, or both, Workday’s algorithm-based hiring tools will be tested in ways it hasn’t been before.
Plaintiffs say the company’s software screened out applicants based on their age, gender, disability, and ethnicity, which violates federal and state employment laws.
The case comes just months after Meta was sued for allegedly discriminating employees during layoffs. The complaint claim that AI-assisted systems disproportionately chose which employees were let go based on who had taken family, medical, or parental leave or who had disabilities.
Given Workday’s customers process about 24% of all U.S. job applications each month, the employment discrimination case may have far-reaching consequences for employers and software companies.
How AI Job Screening Works
Many job applications never reach a human recruiter. Employers often use applicant tracking systems (ATS) to collect resumes, applications and other information. More advanced systems, like that of Workdays, use algorithms and AI to screen, rank and match candidates.
The Equal Employment Opportunity Commission (EEOC) has warned that AI and algorithm-based tools can be used at virtually every stage of employment, from recruiting and screening to interviewing and monitoring. The agency has also noted that applicants may not know when this technology is being used or how decisions are being made, but it’s happening.
Such technology may look for particular skills, qualifications or experience, ask applicants questions through a chatbot, evaluate video interviews and/ or rank candidates. From there, the applicants that filter through the process are reviewed by the employer.
What the Workday Employment Discrimination Lawsuit is About
Mobley v. Workday came after plaintiff Derek Mobley, a Black man over 40 with disabilities, says he repeatedly applied for jobs through companies using Workday’s hiring process and software. He says he was rejected more than 100 times, and claims the software discriminated against him.
Since Mobley filed the Workday lawsuit about hiring bias, the case has quickly expanded and could potentially grow to include millions because they can opt-in to the collective class based on age bias. Plaintiffs argue that the problem of disproportionately discriminating against certain applicants can exist even if the program used was not intentionally programmed to do so.
AI systems can be trained using historical employment data. If that data reflects patters of discrimination or other imbalances, an algorithm can reproduce those patterns while appearing to make neutral decisions.
The lawsuit has reached an important stage. Plaintiffs recently requested the Workday employment discrimination case be certified with four proposed subclasses: African Americans, people over 40, women, and people with disabilities. In response, the U.S. District Judge overseeing the Workday lawsuit scheduled a class-certification hearing for March 9, 2027.
The complaint points to Workday’s own data from an external bias audit involving 724,352 applicants at 10 of its largest customers. Plaintiffs claim that the results showed statistically significant disparities affecting women and African American applicants, as well as those over 40.
Workday denies wrongdoing. The company says its AI tools evaluate job qualifications rather than protected characteristics such as race, age or disability.
Why Proving AI Discrimination is Difficult
While it would be frustrating applying for 100 jobs and receiving 100 rejections, the reasons for those rejections might not be clear.
With a traditional hiring decision, an applicant might eventually realize an interview comment, mistakes on a resume, or other evidence led to not getting the job. The use of algorithms in the hiring process makes this much harder.
What information the program considered, whether it used historical employment data, which qualifications mattered most, if the application was automatically rejected or if a human reviewed it, are possible unknowns.
When the same screening system is used by other employers is another consideration that’s particularly important in the Workday case; plaintiffs allege the common technology used by many companies caused discrimination across hiring decisions, but proving that claim won’t be easy.
Applicants applied for different jobs at different companies. Therefore, they must show that common features of Workday’s program, rather than individual employer decisions, are responsible for the alleged discrimination.
Who’s Liable for Discriminatory Hiring Practices?
Employers who use Workday software are likely to defend themselves by saying they didn’t build the algorithm. On the flip side, Workday’s position is that they didn’t hire anyone – their customers (employers) did.
The EEOC’s position is that software developers or vendors can potentially face liability when their AI or algorithmic tools unlawfully screen out applicants. For the Workday discrimination lawsuit, the agency filed a brief arguing that an AI software vendor may be liable under federal law, specifically, the Age Discrimination in Employment Act (ADEA).
Employers must comply with the ADEA. Its’ protections cover race, sex, national origin, disability and age, regardless of whether a person or technology makes the decision.
But who’s liable gives little comfort for applicants rejected by a biased program. The bigger issue is transparency.
If a computer substantially influences whether someone gets an opportunity to interview, many may want to know if AI was involved in the hiring process and if a human ever saw their application.
AI Lawsuits Surging in the U.S.
Between the Meta layoffs lawsuit and the Workday class action, it’s clear that AI may affect both getting a job and keeping one. But employment lawsuits involving the technology are just one part of a growing wave of litigation.
Authors, news organizations and other copyright owners are suing tech companies over the use of copyrighted material to train AI systems.
Consumers are also filing lawsuits involving chatbot companions and allegations of harm.
Lawmakers and regulators are confronting questions involving AI-generated advertising, consumer information and automated decision-making.
Even the federal government has filed lawsuits. In August, the Justice Department announced a $3.2 million settlement with OpenAI and Statsig over claims that the companies discriminated against U.S. workers during recruitment for certain positions.
Through all of this, the common thread is that AI may still be relatively new, but many of the legal problems it creates are not. Employment discrimination, deceptive business practices, copyright infringement, and personal injury already have legal frameworks.
The difficult question is determining how those laws apply when an algorithm helps make decisions.
