The increasing use of digitized hiring assessments by companies is raising concerns about their impact on disabled workers, according to a recent study. These technologies, which include AI-powered assessments and automated employment decision systems (AEDSs), are intended to streamline the hiring process and reduce bias. However, researchers and advocates warn that they may have discriminatory effects on marginalized groups, particularly individuals with disabilities.
The report titled "Screened Out: The Impact of Digitized Hiring Assessments on Disabled Workers," co-authored by Henry Claypool and Wilneida Negrón, aims to inform various stakeholders about the ethical considerations surrounding these technologies. It focuses specifically on how these assessments affect disabled job applicants.
The study employed a human-centered qualitative approach, engaging participants with disabilities in completing digitized assessments such as personality tests, cognitive tests, and AI-scored video interviews. Participants included individuals with low vision, brain injuries, autism, hearing impairments, intellectual or developmental disabilities, and mobility differences. The research targeted two groups: those seeking hourly jobs and legal professionals pursuing lawyer positions.
Findings indicate that disabled workers often feel discriminated against by these assessments due to accessibility barriers. Contrary to claims by developers that such tools can mitigate bias, participants frequently reported experiencing discrimination. One participant remarked that the tests seemed designed to screen out people with disabilities: “They’re consciously using these tests knowing that people with disabilities aren’t going to do well on them.”
Participants expressed skepticism about whether digitized assessments could be adapted to be fairer or more effective. While some saw potential benefits in remote participation under certain conditions, most doubted the possibility of overcoming inherent biases against disabled individuals within current assessment designs. A participant noted: “We, as very flawed humans, are creating even more flawed tools.”
The study underscores the need for employers and developers to reassess how these technologies are designed and implemented to prevent discrimination against disabled workers. An inclusive approach should consider the diverse needs of all candidates.
Recommendations from the report suggest evaluating the necessity of digitized assessments before implementation and exploring alternative methods that pose less risk of discrimination. If used, assessments should adhere to accessibility guidelines like WCAG or PEAT initiatives and ensure they accurately evaluate relevant skills for specific jobs.
Employers are also encouraged to maintain significant human oversight during assessment processes and use them as supplements rather than replacements for comprehensive evaluations.
