EEOC Recommendations for Bias-Free Use of Decision-Making Technology in Hiring

The Equal Employment Opportunity Commission (EEOC) has recently released guidelines for the use of software, algorithms, and artificial intelligence in the hiring process. These guidelines were issued in response to concerns that this technology could be used to purposefully or inadvertently discriminate against individuals with disabilities, who are protected by the Americans with Disabilities Act (ADA). For more information, we turn to the article, “EEOC Issues Guidance on Employer Use of AI Decision-Making Tools,” written by Julie M. Capell, John P. LeCrone, and Patricia Kinaga of Davis Write Tremaine LLP and shared on Lexology.

The article states that these guidelines are the first substantive output of the EEOC’s Artificial Intelligence and Algorithmic Fairness Initiative, which began in October of 2021 in response to the increased use of such technology by employers. According to the EEOC, the use of tools such as algorithm-based resume scanners, “job fit” testing software, video interviewing software, and virtual assistants has the potential to violate the ADA in the following circumstances:

• The employer relies on algorithmic evaluations of applicants and employees without the provision of reasonable accommodation for individuals with disabilities.
• The decision-making tool screens out qualified candidates with disabilities who could do the job with reasonable accommodation. (Intentionally or unintentionally)
• The tool is used in a manner that goes against the ADA’s restrictions on disability-related inquiries and medical examinations.

With this said, the article shares the following EEOC recommended guidance for using decision-making technology in a manner that is ADA compliant:

1. Train staff to recognize and promptly process requests for reasonable accommodation.

2. Create alternate means of rating job applicants that do not create an unfair disadvantage for qualified candidates with disabilities.

3. Ensure that third party providers of algorithmic decision-making tools either provide reasonable accommodation or promptly forward requests for accommodation to the employer.

4. Use algorithmic decision-making tools that are designed to be accessible to individuals with disabilities.

5. Inform applicants and employees of the availability of reasonable accommodation, and provide clear and accessible instructions for requesting accommodation.

6. Describe the traits that the algorithm is designed to assess and the method by which they are assessed. Describe the variables and factors that may affect one’s rating.

7. Ensure the decision-making tool only measures abilities and qualifications that are truly necessary for the job.

8. Ensure that abilities and qualifications are measured directly, not by correlated characteristics or scores.

9. Ensure that the tool does not ask questions that are likely to elicit information about a disability or seek information about an individual’s physical or mental health.

As a final word of caution, the EEOC states that employers may not rely on a third-party vendor’s claim that a decision-making tool is bias free. Employers are responsible for ensuring that their use of decision-making technology does not discriminate against protected groups. While the use of decision-making technology can make the hiring process efficient and cost effective, it is necessary to do so with caution.