Prominent debates in AI ethics contend that biases are one of the main problems to be eradicated to design fair AI systems, leading the community to develop many technical and non-technical methods mainly focused on eliminating them to create fairer AI tools in healthcare. However, such methods are shown to be often inadequate for the design of fairer healthcare AI. As I argue in this chapter, the problem is a passive understanding of biases as mere errors and technical bugs in AI systems, and of the related solutions mainly aimed at their simple removal. Instead, we should value them as proactive signals of those socio-relational determinants that intangibly and often unfairly shape healthcare access and outcomes in societies over time. This chapter focuses on gender and ethnic biases in healthcare AI and argues for their value in developing AI-based systems and services as drivers of fairer and more just healthcare ecosystems. Also, drawing on ethical theories of affirmative action, the chapter shows how this understanding of biases enables us to generate novel compensatory design actions for the development of truly fair healthcare AI systems, capable of addressing longstanding unfair inequalities in healthcare, therefore promoting better and more just healthcare ecosystems.
Toward Healthcare (Social) Justice: On the Value of Biases in Healthcare AI
Tiribelli, Simona
2025-01-01
Abstract
Prominent debates in AI ethics contend that biases are one of the main problems to be eradicated to design fair AI systems, leading the community to develop many technical and non-technical methods mainly focused on eliminating them to create fairer AI tools in healthcare. However, such methods are shown to be often inadequate for the design of fairer healthcare AI. As I argue in this chapter, the problem is a passive understanding of biases as mere errors and technical bugs in AI systems, and of the related solutions mainly aimed at their simple removal. Instead, we should value them as proactive signals of those socio-relational determinants that intangibly and often unfairly shape healthcare access and outcomes in societies over time. This chapter focuses on gender and ethnic biases in healthcare AI and argues for their value in developing AI-based systems and services as drivers of fairer and more just healthcare ecosystems. Also, drawing on ethical theories of affirmative action, the chapter shows how this understanding of biases enables us to generate novel compensatory design actions for the development of truly fair healthcare AI systems, capable of addressing longstanding unfair inequalities in healthcare, therefore promoting better and more just healthcare ecosystems.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
