AI and Disability: Analyzing Challenges and Embracing Opportunities


Bushra Kundi, McMaster University

The advancement of Artificial Intelligence (AI) in health informatics has a transformative potential for enhancing the health and life experiences of individuals with disabilities. This systematic scoping review delves into the nuanced interplay between disability and AI, scrutinizing both the benefits and challenges of this intersection. The research further utilizes a decolonial theoretical and methodological framework to interrogate the over-reliance on AI systems justified by sustainability and efficiency, rendering issues of accountability and ethical responsibility invalid. . The overarching aim is to articulate how AI aids people with disabilities and to confront the crucial challenges, particularly the pervasive ableist bias within and through AI research, necessitating more decolonial and justice-oriented approaches guiding AI systems development. The theoretical underpinning of the study is informed by the social model. It is rooted in disability justice and decolonial approaches, which center on the disabling effects shaping the everyday realities of disabled people. This framework also calls out the dominance of medical modalities, which predominantly informs current AI research, underscoring the need for a paradigm move away from deficit orientation to diagnosis. The methodological approach adopted in this study follows Arskey and O’Malley’s five-step process, ensuring a comprehensive and systematic exploration of the existing literature. The main arguments presented in the review highlight AIs potential to aid the self-management of health conditions, enhance assistive technologies, and further disability justice. AI has been instrumental in diagnosing conditions like multiple sclerosis and developmental disorders, predicting disease progression, and facilitating rehabilitation. It has also shown promise in developing assistive technologies for communication and mobility and advocating for disability justice by creating platforms for disability advocacy groups. However, the review also unveils critical challenges; it reveals a predominant reliance on the medical model of disability, with a stark underrepresentation of the social model and disability justice in AI research. The articles reviewed demonstrate an acute need for debiasing strategies as a step to decolonize data and AI systems. Specifically, AI models have not sufficiently measured or addressed bias, particularly concerning disabilities. This neglect indicates a broader issue within AI, where ableist perspectives prevail, potentially exacerbating disparities rather than alleviating them. The review emphasizes the need for a collaborative effort to reorient AI development towards disability-justice and decolonial frameworks. It encourages collaboration across disciplines, urging AI technologists to work alongside disability scholars better to incorporate the socio-political and economic aspects of disability into AI systems. Such a transdisciplinary approach promotes empowerment within the disability community, ensuring that AI advancements are leveraged to advance inclusive, accountable, and just technologies beyond neoliberal cost and effect priorities. It calls for transdisciplinary collaborations between AI researchers, disability justice advocates, and scholars to transcend the limitations of the medical model and embrace the broader social and political context of disability. This studys relevance to the Technology and Society session at the CSA Conference lies in its critical analysis of AIs impact on individuals with disabilities—a vital societal issue. It underscores the imperative for AI systems to advance beyond technical excellence to encompass social responsibility and inclusivity. The insights offered call for the creation of AI that upholds social justice, ensuring equitable technological progress that serves the diverse needs of society, especially marginalized social groups. This resonates with the session’s focus on technology’s societal effects, advocating for innovations prioritizing inclusivity and equality.

Non-presenting authors: Fariah Mobeen, York University; Sarah Taleghani, York University; Rachel da Silveira Gorman, York University; Yahya El-Lahib, University of Calgary; Christo El Morr, York University

This paper will be presented at the following session: