Disqualified by Disability: The Exclusion of Disabled Workers by Digitized Hiring Assessments

Paper Title

Disqualified by Disability: The Exclusion of Disabled Workers by Digitized Hiring Assessments

Publication Info

• Topic area: The impact of digitized and AI-based hiring assessments on disabled workers.
• Keywords: AI hiring tools, digitized assessments, disability discrimination, accessibility, employment bias, automated decision systems, human-centered design, emotional toll, legal implications, inclusive hiring.

Background and Problem

• Problem / challenge: Digitized hiring assessments, including AI-based tools, are often inaccessible, discriminatory, and fail to account for the needs of disabled workers. These tools perpetuate biases, create barriers, and may force disclosure of disabilities, leading to exclusion from hiring processes.
• Significance: With 83% of employers and 99% of Fortune 500 companies using automated hiring tools, these systems have a widespread impact on employment opportunities for disabled workers, who already face higher unemployment rates.
• Motivation and related work: Prior research highlights the risks of bias in AI hiring tools, including racial, gender, and disability discrimination. However, there is limited understanding of the specific experiences of disabled workers with digitized assessments, particularly across diverse disabilities. This paper addresses this gap by focusing on the perspectives of disabled U.S.-based workers.

Solution

• Proposed approach: A qualitative, human-centered study examining the experiences of disabled workers with digitized hiring assessments, including personality tests, cognitive tests, gamified assessments, and AI-scored video interviews.
• Novelty:
  1. Focus on the firsthand experiences of disabled workers across diverse disabilities.
  2. Examination of accessibility barriers and emotional toll in digitized assessments.
  3. Analysis of the discriminatory nature of AI-based hiring tools.
  4. Development of guidelines for reducing harm and improving inclusivity in hiring technologies.
• Procedure and key techniques:
  • Participants (n=17) with diverse disabilities completed simulated hiring assessments on two commercial platforms.
  • Semi-structured interviews captured their reflections on accessibility, emotional impact, and perceptions of fairness.
  • Thematic analysis identified key barriers, emotional costs, and recommendations for improvement.

Results

• Concrete findings:
  • 53% of participants were unable to complete at least one assessment due to inaccessibility.
  • 41% scored below the 5th percentile on at least one test.
  • Emotional intelligence tests and gamified assessments were particularly inaccessible and ineffective for disabled participants.
  • Automated facial analysis in video interviews misclassified emotions, often labeling disabled participants as "neutral," "angry," or "disgusted."
• Advantage over baselines: The study provides qualitative insights into the lived experiences of disabled workers, highlighting systemic flaws in hiring assessments that are not addressed in prior quantitative studies.
• Experiments / evaluation:
  • Two 2-hour assessment sessions followed by a 1-hour semi-structured interview.
  • Participants included hourly workers and attorneys, representing a range of educational and professional backgrounds.
  • Analysis focused on accessibility barriers, emotional toll, and discriminatory outcomes.
• Limitations and future work:
  • Findings are not generalizable due to the small, qualitative sample.
  • Future work should include larger, more diverse samples and examine real-world hiring processes.
  • Further research is needed to evaluate the effectiveness and fairness of specific assessment tools.

Summary

This study investigates the discriminatory impact of digitized hiring assessments on disabled workers, revealing significant accessibility barriers, emotional tolls, and systemic biases. Participants reported difficulties completing tests, low scores unrelated to job qualifications, and forced disclosure of disabilities. The findings challenge claims that AI hiring tools reduce bias and highlight the need for transparency, human oversight, and supplementary use of assessments. The study provides actionable guidelines for developers and employers to create more inclusive hiring technologies, emphasizing the importance of accessibility and fairness in employment practices.