May 2025
Legal Risks of Employment Discrimination for AI Recruitment Service Providers: A Case Study of Mobley v. Workday in the U.S.
In July 2024, the United States District Court for the Northern District of California issued a ruling in Mobley v. Workday, Inc., 740 F.Supp.3d 796 (2024), holding that a third-party AI recruitment software vendor could also be named as a defendant in employment discrimination lawsuits as an employer in the U.S. This case differs from typical employment discrimination disputes arising from employers' use of AI recruitment tools and carries instructive significance given the increasing prevalence of third-party AI recruitment services in Taiwan. The following outlines the facts, issues, court’s holdings and reasoning, and implications:
1. Case Summary
The plaintiff, Derek Mobley, is a 40-year-old African American male who suffers from anxiety and depression. He filed a lawsuit against the defendant, Workday, Inc., an AI recruitment software vendor, alleging that Workday’s algorithmic decision-making tools discriminate against him and job applicants who are African American, over the age of forty, and/or disabled. He asserts U.S. federal law claims under Title VII of the Civil Rights Act of 1964, the Age Discrimination in Employment Act of 1967, and the Americans with Disabilities Act Amendments Act of 2008 (collectively referred to as “the U.S. federal employment discrimination laws”).
Mobley claimed that he had been rejected from over 100 positions across different industries and employers. These positions all used the AI recruitment tools provided by Workday. Mobley believed that Workday's AI recruitment tools discriminated against African Americans with disabilities who were over 40 years old, like himself. Workday argued that it was merely an AI recruitment software vendor, not an employer as defined by the U.S. federal employment discrimination laws, and therefore should not be separately held liable for functions performed on behalf of employers. Besides, Workday contended that Mobley failed to plead a prima facie case of disparate impact and therefore filed a motion to dismiss the disparate impact claims.
2. Main Issues of This Case
(1) Whether Workday qualifies as an employer under the U.S. federal employment discrimination laws.
(2) Whether Mobley has pleaded a prima facie case that Workday’s AI recruitment tools caused the disparate impact on him.
3. Court’s Holdings and Reasoning
(1) Workday qualifies as an employer under the U.S. federal employment discrimination laws.
The definition of "employer" in the U.S. federal employment discrimination laws includes any person acting as an agent of an employer. Under the interpretation of these laws, a third-party agent may be liable as an employer where the agent has been delegated functions traditionally exercised by an employer. Otherwise, when a third party intentionally sells software to employers that may constitute employment discrimination, but the employers themselves are unaware, no one would be held responsible for intentional discrimination.
The AI recruitment tools provided by Workday do not simply implement in a rote way the criteria that employer customers set forth, but integrate with artificial intelligence and machine learning to recommend or reject candidates in the hiring decision-making process. These features of the tools are at the core of traditional employment functions and at the heart of equal access to employment opportunities that the U.S. federal employment discrimination laws seek to protect.
Therefore, Workday is not merely a software vendor but acts as an agent performing a traditional hiring function of an employer, meeting the definition of an employer under the U.S. federal employment discrimination laws. Otherwise, employers could delegate employment-related decisions to third-party algorithmic decision-making tools that are beyond the reach of the U.S. federal employment discrimination laws, making it difficult for employees and applicants to challenge employment discrimination.
(2) Mobley has pleaded a prima facie case that Workday’s AI recruitment tools caused a disparate impact on him.
To plead a prima facie case of disparate impact discrimination, a plaintiff must (a) show a significant disparate impact on a protected class or group; (b) identify the specific employment practices or screening criteria being challenged; and (c) show a causal relationship between the challenged employment practice or criteria and the disparate impact.
Workday embeds artificial intelligence and machine learning into its AI recruitment tools to screen whether the applicants meet job qualifications. The tools can also, at the employer customers’ hiring preferences, have applicants take personality tests or pymetrics tests that use neuroscience data and AI 1 . The tools will determine whether an applicant should be recommended or rejected based on the resumes or information provided by applicants and data obtained from the tests.
In this case, Mobley plausibly alleges that Workday's AI recruitment tools rely on biased training data and that applicants with mental health issues or cognitive disorders perform worse on the tests used by the tools to gather information. Although not all employer customers use these tests, Mobley was rejected from over 100 positions across different industries and employers. Combined with Mobley's plausible allegations, it plausibly supports an inference that Workday’s algorithmic tools disproportionately reject applicants based on factors other than qualifications, such as race, age, or disability. It is sufficient to allege a disparate impact on Mobley’s protected traits and to identify the specific employment practice being challenged at this stage.
Furthermore, Mobley was rejected from over 100 positions even though he was allegedly qualified for the positions. He received rejection emails outside regular business hours, sometimes less than an hour after applying. These circumstances plausibly support that Mobley was automatically rejected by Workday based on factors other than qualifications, such as protected traits. Therefore, Mobley has pleaded a prima facie case of disparate impact caused by Workday’s AI recruitment tools.
4. Implications of the Case and Subsequent Developments
In Taiwan, although the Employment Service Act does not explicitly define "employer" to include anyone acting as an agent of an employer, as in the U.S. federal employment discrimination laws, there has been a court decision that references Article 2, Subparagraph 2 of the Labor Standards Act to determine that the term "employer" under Article 2, Subparagraph 3 of the Employment Service Act includes the person who represents the business owner in handling workers’ matters 2 . This means that third-party AI interview and recruitment service providers in Taiwan could potentially be deemed as persons performing hiring functions on behalf of an employer, and even be held separately liable for employment discrimination.
This case is currently being heard in the United States District Court for the Northern District of California. At this stage, the court has only determined that a third-party AI recruitment software vendor qualifies as an employer under the U.S. federal employment discrimination laws. Whether AI recruitment tools that integrate AI and machine learning constitute employment discrimination is still under review. The court has recently granted preliminary collective certification for Mobley to notify similarly situated individuals of this case and provide them an opportunity to opt-in to having their claims heard on a collective basis 3 . This case is expected to develop into a landmark case.
This case has only preliminarily resolved a procedural issue, that is, a third-party AI recruitment software vendor qualifies as a defendant under the U.S. federal employment discrimination laws. Whether algorithmic recruitment decision-making tools are discriminatory deserves continued attention.
1. The pymetrics test refers to the service originally provided by Pymetrics, a company founded in New York. The company claimed that it could assess job applicants' cognitive and emotional traits, including risk tolerance and how quickly the applicant responds to situations, within 25 minutes through a series of questions or games. The company said its system ensured more fairness compared to simply screening an applicant’s resume or CV. Please see: Andrea Murad, The Computers Rejecting Your Job Application, BBC (Feb. 6, 2021): https://www.bbc.com/news/business-55932977. Pymetrics was acquired by Harver in August 2022. Please see: Harver Acquires Pymetrics, Further Enhancing Talent Decision Capabilities Across the Employee Lifecycle, Harver: https://harver.com/press/harver-acquires-pymetrics/ (last visited May 20, 2025).
2. The 104-Jian-Zi-24 Administrative Decision of the Taiwan New Taipei District Court.
3. Mobley v. Workday, Inc., No. 3:23-cv-00770 (N.D. Cal. May 16, 2025) (order granting preliminary collective certification).
1. Case Summary
The plaintiff, Derek Mobley, is a 40-year-old African American male who suffers from anxiety and depression. He filed a lawsuit against the defendant, Workday, Inc., an AI recruitment software vendor, alleging that Workday’s algorithmic decision-making tools discriminate against him and job applicants who are African American, over the age of forty, and/or disabled. He asserts U.S. federal law claims under Title VII of the Civil Rights Act of 1964, the Age Discrimination in Employment Act of 1967, and the Americans with Disabilities Act Amendments Act of 2008 (collectively referred to as “the U.S. federal employment discrimination laws”).
Mobley claimed that he had been rejected from over 100 positions across different industries and employers. These positions all used the AI recruitment tools provided by Workday. Mobley believed that Workday's AI recruitment tools discriminated against African Americans with disabilities who were over 40 years old, like himself. Workday argued that it was merely an AI recruitment software vendor, not an employer as defined by the U.S. federal employment discrimination laws, and therefore should not be separately held liable for functions performed on behalf of employers. Besides, Workday contended that Mobley failed to plead a prima facie case of disparate impact and therefore filed a motion to dismiss the disparate impact claims.
2. Main Issues of This Case
(1) Whether Workday qualifies as an employer under the U.S. federal employment discrimination laws.
(2) Whether Mobley has pleaded a prima facie case that Workday’s AI recruitment tools caused the disparate impact on him.
3. Court’s Holdings and Reasoning
(1) Workday qualifies as an employer under the U.S. federal employment discrimination laws.
The definition of "employer" in the U.S. federal employment discrimination laws includes any person acting as an agent of an employer. Under the interpretation of these laws, a third-party agent may be liable as an employer where the agent has been delegated functions traditionally exercised by an employer. Otherwise, when a third party intentionally sells software to employers that may constitute employment discrimination, but the employers themselves are unaware, no one would be held responsible for intentional discrimination.
The AI recruitment tools provided by Workday do not simply implement in a rote way the criteria that employer customers set forth, but integrate with artificial intelligence and machine learning to recommend or reject candidates in the hiring decision-making process. These features of the tools are at the core of traditional employment functions and at the heart of equal access to employment opportunities that the U.S. federal employment discrimination laws seek to protect.
Therefore, Workday is not merely a software vendor but acts as an agent performing a traditional hiring function of an employer, meeting the definition of an employer under the U.S. federal employment discrimination laws. Otherwise, employers could delegate employment-related decisions to third-party algorithmic decision-making tools that are beyond the reach of the U.S. federal employment discrimination laws, making it difficult for employees and applicants to challenge employment discrimination.
(2) Mobley has pleaded a prima facie case that Workday’s AI recruitment tools caused a disparate impact on him.
To plead a prima facie case of disparate impact discrimination, a plaintiff must (a) show a significant disparate impact on a protected class or group; (b) identify the specific employment practices or screening criteria being challenged; and (c) show a causal relationship between the challenged employment practice or criteria and the disparate impact.
Workday embeds artificial intelligence and machine learning into its AI recruitment tools to screen whether the applicants meet job qualifications. The tools can also, at the employer customers’ hiring preferences, have applicants take personality tests or pymetrics tests that use neuroscience data and AI 1 . The tools will determine whether an applicant should be recommended or rejected based on the resumes or information provided by applicants and data obtained from the tests.
In this case, Mobley plausibly alleges that Workday's AI recruitment tools rely on biased training data and that applicants with mental health issues or cognitive disorders perform worse on the tests used by the tools to gather information. Although not all employer customers use these tests, Mobley was rejected from over 100 positions across different industries and employers. Combined with Mobley's plausible allegations, it plausibly supports an inference that Workday’s algorithmic tools disproportionately reject applicants based on factors other than qualifications, such as race, age, or disability. It is sufficient to allege a disparate impact on Mobley’s protected traits and to identify the specific employment practice being challenged at this stage.
Furthermore, Mobley was rejected from over 100 positions even though he was allegedly qualified for the positions. He received rejection emails outside regular business hours, sometimes less than an hour after applying. These circumstances plausibly support that Mobley was automatically rejected by Workday based on factors other than qualifications, such as protected traits. Therefore, Mobley has pleaded a prima facie case of disparate impact caused by Workday’s AI recruitment tools.
4. Implications of the Case and Subsequent Developments
In Taiwan, although the Employment Service Act does not explicitly define "employer" to include anyone acting as an agent of an employer, as in the U.S. federal employment discrimination laws, there has been a court decision that references Article 2, Subparagraph 2 of the Labor Standards Act to determine that the term "employer" under Article 2, Subparagraph 3 of the Employment Service Act includes the person who represents the business owner in handling workers’ matters 2 . This means that third-party AI interview and recruitment service providers in Taiwan could potentially be deemed as persons performing hiring functions on behalf of an employer, and even be held separately liable for employment discrimination.
This case is currently being heard in the United States District Court for the Northern District of California. At this stage, the court has only determined that a third-party AI recruitment software vendor qualifies as an employer under the U.S. federal employment discrimination laws. Whether AI recruitment tools that integrate AI and machine learning constitute employment discrimination is still under review. The court has recently granted preliminary collective certification for Mobley to notify similarly situated individuals of this case and provide them an opportunity to opt-in to having their claims heard on a collective basis 3 . This case is expected to develop into a landmark case.
This case has only preliminarily resolved a procedural issue, that is, a third-party AI recruitment software vendor qualifies as a defendant under the U.S. federal employment discrimination laws. Whether algorithmic recruitment decision-making tools are discriminatory deserves continued attention.
1. The pymetrics test refers to the service originally provided by Pymetrics, a company founded in New York. The company claimed that it could assess job applicants' cognitive and emotional traits, including risk tolerance and how quickly the applicant responds to situations, within 25 minutes through a series of questions or games. The company said its system ensured more fairness compared to simply screening an applicant’s resume or CV. Please see: Andrea Murad, The Computers Rejecting Your Job Application, BBC (Feb. 6, 2021): https://www.bbc.com/news/business-55932977. Pymetrics was acquired by Harver in August 2022. Please see: Harver Acquires Pymetrics, Further Enhancing Talent Decision Capabilities Across the Employee Lifecycle, Harver: https://harver.com/press/harver-acquires-pymetrics/ (last visited May 20, 2025).
2. The 104-Jian-Zi-24 Administrative Decision of the Taiwan New Taipei District Court.
3. Mobley v. Workday, Inc., No. 3:23-cv-00770 (N.D. Cal. May 16, 2025) (order granting preliminary collective certification).
