EEOC Warns of ADA Violation Potential From AI Tools
Author: Emily Scace, XpertHR Legal Editor
May 16, 2022
Using artificial intelligence (AI) and other software tools to make employment decisions may lead to disability discrimination in violation of the Americans with Disabilities Act (ADA), according to new guidance from the Equal Employment Opportunity Commission (EEOC).
AI and other software tools are increasingly a part of many employers' hiring and screening processes, says the EEOC. For example, computer software may score the resumes of applicants for a particular job, or candidates may be asked to take a computer-based test. But while these processes may save time and resources in recruiting and hiring, they have the potential to disadvantage or exclude individuals with disabilities if implemented without proper safeguards.
According to the EEOC, the most common ways that an employer's use of AI or other software tools could violate the ADA are:
- Failing to provide a reasonable accommodation that is necessary for a job applicant or employee to be rated fairly and accurately;
- Relying on a tool that intentionally or unintentionally screens out an individual with a disability, even though the individual is able to do the job with or without a reasonable accommodation; and
- Adopting a tool that violates ADA restrictions on disability-related inquires and medical examinations.
The guidance specifies that even if the AI tool or software is designed or administered by an entity other than the employer, such as a software vendor, the employer may still be liable for any discriminatory results it produces. In addition, the EEOC cautions employers not to rely on vendor or provider claims that a particular tool is "bias-free." Even if a provider has taken steps to prevent or reduce disability bias in its tools, each disability is unique, the EEOC notes, and measures aimed at preventing some types of disability discrimination will not necessarily prevent the tool from disadvantaging an individual with a different disability.
Steps an employer using AI or algorithmic decision-making tools can take to reduce the chances of violating the ADA include:
- Clearly indicating that reasonable accommodations are available to people with disabilities and providing clear instructions for requesting an accommodation;
- Training staff to recognize and process requests for reasonable accommodations as quickly as possible;
- Allowing accommodations such as extended time or alternative versions of tests or assessments for individuals with a disability-related need;
- Ensuring that materials presented to job applicants and employees are available in alternative formats;
- Only developing and selecting tools that measure abilities and qualifications that are truly necessary for a particular job; and
- Providing job applicants and employees who will be assessed by an algorithmic decision-making tool with as much information about the tool as possible, including:
- Traits or characteristics it is designed to measure;
- Methods for measuring those traits or characteristics; and
- The disabilities, if any, that might potentially lower the assessment results.