Abigail Murphy

From Fairness to Justice: Reframing Ethical AI in Disability Diagnosis

Introduction

From fairness to justice: reframing ethical ai in disability diagnosis . Reframes ethical AI in disability diagnosis, shifting from fairness to justice. Explores power dynamics & systemic issues to build responsible, equitable AI systems for disabled individuals.

Abstract

In recent years, the use of artificial intelligence (AI) as a potential tool for diagnosing disabilities through various machine learning (ML) processes has gained significant attention in scientific, medical, and philosophical realms. ​The current discourse on equitable AI use, specifically for disability diagnosis, primarily revolves around a fairness-centred approach, seeking to mitigate bias and ensure fairness in AI design and development (Trewin, 2018, p. 2).  The current focus on fairness does not adequately identify and address the systemic frameworks that place disabled individuals at a disadvantage within our society. By evaluating the ethical implications of AI use for disabled people under the narrow scope of “fairness,” we risk legitimizing harmful power dynamics and pre-existing inequalities for disabled individuals and other marginalized groups. ​This essay explores the limitations of fairness and the potential of justice as a guiding principle for responsible and ethical AI in the context of disability. By exploring power dynamics, epistemic authority, and potential objections, I aim to underscore the need for reframing from fairness to justice in addressing the complex ethical implications of AI for individuals with disabilities.  ​Drawing on the work of feminist scholars, I argue that justice, as opposed to fairness, can drive recovery, remedy past harm, and promote more responsible and ethical AI systems. 

Review

This essay proposes a timely and critical intervention in the discourse surrounding ethical AI, particularly within the sensitive domain of disability diagnosis. By meticulously dissecting the limitations of a predominantly "fairness-centered" approach, the author argues compellingly that such a framework is insufficient to address the deep-seated systemic disadvantages faced by disabled individuals. The paper's central contribution lies in advocating for a profound conceptual shift towards "justice" as a more encompassing and transformative guiding principle. This reframing is posited as essential to prevent the legitimization of existing inequalities and harmful power dynamics, thereby advancing a more truly ethical and responsible integration of AI in healthcare. The paper's strength lies in its incisive critique and its ambitious proposal for a more robust ethical foundation. By engaging with feminist scholarship, the author effectively highlights how a narrow focus on fairness can inadvertently perpetuate or mask deeper structural issues, particularly in areas concerning marginalized groups. The proposed exploration of power dynamics and epistemic authority is particularly pertinent, offering a lens through which to interrogate how AI might replicate or challenge existing hierarchies in diagnostic processes. The vision that justice can actively drive recovery and remedy past harm presents a proactive and transformative goal for AI development, moving beyond mere bias mitigation towards genuine societal equity and empowerment for individuals with disabilities. To further solidify its impact, the paper could benefit from a more detailed exploration of the practical implications of a justice-centered AI. While the theoretical argument is robust, concrete examples illustrating how a justice framework would specifically alter AI design, data collection, or diagnostic interpretation—contrasted with a fairness-based approach—would significantly enhance its actionable value. Additionally, a discussion of the potential challenges in operationalizing "justice" within technical AI development, including the development of new metrics or methodologies to assess justice, would provide invaluable guidance for researchers and practitioners striving to implement these critical ethical principles.

Full Text

Comments

You need to be logged in to post a comment.