Over the past several years, workplace artificial intelligence (“AI”) tools have matured from novel to mainstream.  Whether facilitating attracting, screening, hiring, and onboarding job applicants or charting the career path or promotability of current employees, workplace AI tools will likely become more prevalent.  Legislators and administrative agencies have taken note and are in various stages of examining and regulating these tools, with the primary goal of ensuring that they do not violate federal and state laws prohibiting workplace discrimination.

The Equal Employment Opportunity Commission (EEOC) is one of the federal agencies that is keenly aware of and has raised concerns about the potential of automated technologies to perpetuate the biases that humans may intentionally or unintentionally inject into hiring and employment decisions.  In October 2021, the EEOC launched the Artificial Intelligence and Algorithmic Fairness Initiative in response to the increased use of AI in workforce management systems and hiring processes.  According to the EEOC, the purpose of the initiative is “to ensure that the use of software, including [AI], machine learning and other emerging technologies used in hiring and other employment decisions comply with the federal civil rights laws that the EEOC enforces.”

Consistent with that initiative, on May 12, 2022, the EEOC issued its first guidance addressing the use of AI in making employment decisions.  The guidance, entitled “The Americans with Disabilities Act and the Use of Software, Algorithms, and Artificial Intelligence to Assess Job Applicants and Employees,” aims to ensure that AI and other automated technologies do not discriminate against or disadvantage job applicants and employees with disabilities.

The EEOC’s non-binding, technical guidance provides what it believes to be “best practices” for employers to reduce the likelihood of disability discrimination in their use of AI technologies.  AI tools such as résumé scanners, virtual assistants/chatbots, and video-interviewing, personality, aptitude, and skill software, are the areas of primary interest.  The guidance outlines three general areas of concern in the context of the Americans with Disabilities Act (ADA) for employers using algorithmic decision-making tools:

1. Employers should have a process to provide reasonable accommodations for job seekers being evaluated by such tools;

2. Employers should be careful that workers with disabilities evaluated by algorithmic decision-making tools are not “screened out” from consideration in a job or promotion where they can do the job with or without a reasonable accommodation; and

3. The use of algorithmic decision-making tools that cause applicants or employees to provide information about disabilities or medical conditions must not result in prohibited disability-related inquiries or medical exams.

 The EEOC’s technical guidance also includes a section on “Promising Practices for Employers,” which includes advice such as:

• using tools that have been designed to be accessible to individuals with as many different kinds of disabilities as possible;

• informing all candidates that reasonable accommodations are available for individuals with disabilities, and providing clear and accessible instructions for requesting such accommodations;

• describing, in plain language and accessible formats, the traits the tech-enabled tool is designed to assess, the method by which those traits will be assessed, and the variables or factors that may affect the assessment or rating;

• training staff to recognize and process requests for reasonable accommodation as quickly as possible, and to develop or obtain alternative means of assessing candidates when the current procedure is inaccessible or otherwise unfairly disadvantages someone who has requested a reasonable accommodation because of a disability;

• if working with a vendor, devising a process such that will allow the vendor to quickly relay any requests for reasonable accommodation to the employer;

• ensuring that the tool measures only abilities or qualifications that are truly necessary for the job, and those abilities and qualifications are measured directly, rather than by way of characteristics that are correlated with those abilities or qualifications; and

• confirming that the tool does not ask candidates questions that are likely to elicit information about a disability or seek information about an individual’s physical or mental impairments or health, unless such inquiries are related to a request for a reasonable accommodation.

Preventing employers from using tools that unfairly “screen out” workers with disabilities is the guidance’s key focus.  According to the EEOC, “screen out” occurs “when a disability prevents a job applicant or employee from meeting—or lowers their performance on—a selection criterion, and the applicant or employee loses a job opportunity as a result.”  By way of example, the EEOC cites video interviewing software that analyzes applicants’ speech patterns, but fails to fairly or accurately score an applicant with a speech impediment.

Incorporating familiar guidance, the EEOC also instructs employers not to use tools that may directly or indirectly elicit information about a disability, or that directly ask an applicant or employee whether the individual has a disability.  Going one step farther, however, the EEOC cautions employers that even if a tool does not violate the ADA’s restrictions on disability-related inquiries and medical examinations, it still might violate other parts of the ADA.  For instance, “if a personality test asks questions about optimism, and if someone with Major Depressive Disorder (“MDD”) answers those questions negatively and loses an employment opportunity as a result, the test may ‘screen out’ the applicant because of MDD.”  Such a “screen out,” according to the EEOC, may be unlawful if the applicant can perform the essential functions of the job, with or without reasonable accommodation.

What Should Employers Do Now?

Employers using or considering using an algorithmic decision-making tool should carefully study this recent EEOC guidance, which provides insight into how the EEOC will evaluate AI tools.  Compliance with these guidelines will minimize potential risk for claims of disability discrimination when using algorithmic decision-making technologies for use in hiring and other employment decisions.

The EEOC’s guidance, however, does have limitations.  For instance, the guidance suggests that employers provide sufficient notice to applicants about the metrics measured by the tool, how the tool measures them, and the variables or factors that might lower an assessment, such that a user with a disability would know to ask for a reasonable accommodation.  While a laudable goal, it may not be possible for an employer to succinctly identify every variable or factor that could impact an assessment.  Similarly, the EEOC’s recommendation that employers use tools that have been designed to be accessible to individuals with as many kinds of disabilities may not be technologically feasible.  Therefore, to address both issues, employers should provide clear notice to applicants about the availability of reasonable accommodations, be prepared to engage in the interactive process, and have alternative assessment procedures in place for those whom the tool cannot accurately measure or assess.

In addition, employers should consult with their vendors who provide and administer these tools to identify any potential for discriminatory impact and evaluate their accommodations processes for both applicants and employees.  Employers may also wish to consider seeking a warranty (or indemnification) that a vendor’s product does not discriminate against individuals with disabilities (or other protected classes).