Artificial intelligence can make recruiting faster, but legal responsibility does not disappear when technology enters the hiring process. That point is being tested in Mobley v. Workday, Inc., Case No. 3:23-cv-00770- RFL (N.D. Cal. 2023), a federal lawsuit pending in the Northern District of California, in which job applicants allege that Workday’s AI-based applicant-screening tools discriminated on the basis of race, age and disability. The court has allowed key claims to proceed and granted preliminary certification of a nationwide collective action under the Age Discrimination in Employment Act (ADEA). Those rulings do not establish that discrimination occurred, but the case illustrates the legal scrutiny that can follow when automated tools influence employment decisions.
AI can help employers draft job descriptions, source applicants, rank resumes and administer assessments. But as employers incorporate these tools into recruiting, they still need to understand what the technology evaluates, how it affects candidates and where human judgment must enter the process.
AI Does Not Change the Rules
The federal government has reacted to the use of AI in the employment context with fits and starts, with the EEOC issuing guidance in 2023 and then discreetly removing it from its website in 2025. Nevertheless, federal anti-discrimination laws remain unchanged, and the same rules that apply to an employer’s decision also apply when technology helps source, screen, rank, assess or reject applicants. Title VII still prohibits discrimination based on race, color, religion, sex and national origin; the ADEA still protects applicants age 40 and older; and the Americans with Disabilities Act still prohibits disability discrimination and requires reasonable accommodations in the application process. An AI tool may violate Title VII if it uses a protected trait (or a neutral proxy for a protected trait such as a name, ZIP code or alma mater) to make recommendations or screen out resumes. The use of automated assessments could screen out qualified individuals with disabilities who may need reasonable accommodation.
Furthermore, purchasing a tool from a vendor does not absolve the employer of liability for the effects of that tool. Existing discrimination laws can cover both intentional discrimination and, in appropriate circumstances, unintentional disparate impact. Therefore, all HR professionals must ensure that their hiring process treats all applicants lawfully. Employers should make sure that everyone using a tool (whether in HR or a hiring manager) understands the factors it evaluates, ensures that there is a process for accommodations and includes a step for human review. Employers should also monitor whether the tool’s outcomes reveal patterns of disproportionate exclusion. In short, AI should be integrated into an employer’s existing equal-employment-opportunity processes.
State and Local Rules Matter
In addition to remaining mindful of federal employment laws, employers (especially those with multi-state workforces) must monitor state-specific laws related to AI use in the employment arena. This means staying abreast of legal changes in every state where an employer has employees (including those who are remote), not just where the employer is headquartered. Some states and local jurisdictions have specifically passed laws regarding AI use in the employment context. For example, New York City prohibits employers from using automated employment decision tools in making employment decisions unless the employer (1) conducts an annual bias audit of the tool; (2) publishes a public summary of that audit and (3) provides certain notices to employees and applicants. Similarly, Illinois’ Artificial Intelligence Interview Act requires employers who use AI to record or analyze interviews to inform applicants of the AI use, provide a written explanation of how the tool works and get the applicant’s consent. Further, California’s Civil Rights Department has issued regulations clarifying that automated-decision tools may violate the state’s anti-discrimination statutes and require employers to maintain automated decision data for a minimum of four years. State law on this front may change rapidly, so it is important to stay current.
Five Practical Safeguards
No single checklist can eliminate every risk, but employers can reduce exposure by adopting several basic controls.
- Inventory all AI use. Inventory all AI tools that touch on your recruiting/hiring process in some way. This could include obvious things like AI resume screening or less noticeable items, such as a tool that records all meeting notes but is also used in virtual interviews.
- Conduct meaningful vendor diligence. Before using a new AI tool, ask whether the vendor has completed a bias audit or validity study, how it measures selection rates by demographic group, who selected and weighed the factors, how frequently the algorithm changes and whether the tool collects biometric information. Confirm how the vendor stores, safeguards and deletes applicant data. Check if it has been the subject of any litigation.
- Build accessibility into the process. Chatbots that reject candidates with employment gaps may unlawfully screen out applicants whose gaps relate to a disability. Likewise, AI that evaluates speech patterns may disadvantage an applicant with a speech impairment. Make sure you provide a clear accommodation-request process and alternative assessments where appropriate.
- Keep a human in the loop. Do not permit a system to make an unreviewed final rejection or hiring decision. Someone in HR or a properly trained hiring manager should also question the recommendation, make the final hiring decision and document the basis for the decision.
- Test, monitor and document. Before a broad rollout, test any new AI tools in a limited setting. Establish success metrics, evaluate outcomes through bias audits and validity studies and prepare a response plan if disparities emerge.
AI can be a useful part of recruiting, but a desire for efficiency does not eliminate an employer’s responsibility to administer a lawful hiring process. The practical lesson is not that employers should avoid these tools, but that they should remain informed and accountable for how the tools operate within their recruiting practices.
Mobley demonstrates that when an automated hiring process is challenged, the technology itself can become part of the legal inquiry, along with how it was selected and used and what oversight surrounded the resulting decisions. Employers that understand their AI tools and incorporate appropriate compliance measures before problems arise will be better positioned to take advantage of the technology while managing the legal risks that come with it.
Ashley Kelly is a partner at Arnall Golden Gregory LLP and leader of the firm’s employment practice group. She serves as lead counsel in a full range of employment and business disputes, including claims involving allegations of discrimination and harassment, violations of employee leave laws, wage and hour issues, breach of executive employment contracts, benefits disputes, and other complex business litigation matters. She also negotiates executive compensation packages and advises on matters relating to unfair competition and restrictive covenants. She can be reached at ashley.kelly@agg.com.
Lindsey Locke is an associate at Arnall Golden Gregory LLP and a member of the firm’s employment practice. She advises employers on a range of employment matters, including employment disputes, workplace policies and compliance, restrictive covenants, employee leave and accommodations, and discrimination and other workplace issues. She also represents clients in employment-related litigation and provides guidance on developing legal and practical strategies for navigating changes in federal and state employment laws. She can be reached at lindsey.locke@agg.com.