Human-Centered AI for Accessibility: Designing Transparent Intelligent Systems for the Disabled Workforce

Abstract

The given quantitative study investigates the way human-oriented and accessible AI-enhanced features aid disabled workers at work. Based on survey information on 120 respondents, the study quantifies four key areas including trust and understanding due to explainability, task efficiency due to multimodal interaction, perceived inclusion and fairness, and organizational readiness. The findings are that 78% of the users have more trust in AI when explanations are available, the task time can be improved up to 32 mode by multimodal features, and that available AI devices make individuals feel included with 68% of respondents. Nevertheless, there is a lack of organizational support. The report points out that readily available AI enhances equity, efficiency, and independence of the user.