In the blind spot of AI-based mobility: eye-tracking visual attention to people with reduced mobility in autonomous vehicles and the ethics of inclusive design

As AI systems increasingly shape urban mobility futures, the principles of equity, diversity, and inclusion have become central to responsible design. While race, gender, and age have been extensively studied, people with reduced mobility (PRM) remain underrepresented in empirical research on AI and autonomous vehicles (AVs). This article addresses this critical gap by investigating how disability and intersecting social categories are perceived and attended to in the visual framing of AV from an ethical standpoint. Our aim is to understand how people cognitively perceive PRM in automated decision-making contexts, and what this reveals about inclusive AI-based mobility design. We conducted a large-scale online eye-tracking experiment (N = 1272) combined with a survey to assess visual attention toward PRM in AV. Results reveal that PRM are consistently overlooked in visual attention compared to other demographic categories. Cluster analysis identified three distinct profiles of perception: High Disability Awareness, Adaptive Disability Evaluators, and Low Disability and Diversity Awareness, the latter encompassing most respondents. These insights were validated in an in-lab experiment with data science experts. Despite broad endorsement of fairness, justice, and diversity as guiding values in AV development, actual gaze behavior suggests a disconnect between normative commitments and perceptual engagement. We argue that this “blind spot” in AI-based mobility reflects deeper ableist assumptions embedded in algorithmic perceptions and raises ethical concerns. To counter this, we call for inclusive, cross-sectoral advisory boards that include PRM communities, ensuring that algorithmic designs not only reflect diversity but ethically see it.