Ethical design & AI for Health: A structured review and theoretical critique

ABSTRACT:

Emerging digital technologies such as artificial intelligence (AI) not only mediate everyday human practices and engagement with the world, but are also rapidly re-creating new social and economic orders that characterize contemporary global life. The speed at which these changes have occurred, and the lack of adequate response at either the national or global policy level, has inspired renewed interest in the literature on ethics and values in design (E+VID). These “ethics-first” approaches generally focus on clarifying the normative dimensions of design, and outlining strategies to more strongly incorporate ethical decision-making throughout the design process. Only recently, however, has attention been devoted to comparing these approaches on their practical and normative characteristics, and very little attention has been paid to their application in health-related AI. The purpose of this session is to present a review of E+VID approaches that have appeared in the academic literature, underscoring in particular the central importance of attending to designer agency and normative strength for such approaches to be meaningfully employed in the context of AI for health.

Our approach in this project was framed by a highly specific objective: to identify approaches to E+VID that have been proposed and discussed in the literature, and to assess those approaches for their normative strength and assumptions about designer agency. In this way, our approach to identifying relevant publications was informed by purposive sampling. This process uncovered 17 E+VID approaches, ranging from normatively “weak” approaches to normatively “strong” approaches. Normatively weaker approaches included ‘value-sensitive design’, which emphasizes identifying and incorporating the values of the local design team throughout the design process. Normatively stronger approaches include ‘design justice’, which is concerned with how the design process distributes risks, harms, and benefits among different groups in society. 

We conclude by discussing the strengths and limitations of normatively “strong” versus “weak” approaches, and suggest that design ethics in health care must consider a particularly complex range of institutional realities that surround the practice of health-related design.

AUTHOR BIO: 

Joseph Donia is a current master's student, and incoming PhD student at the Institute for Health Policy, Management & Evaluation at the University of Toronto. In 2018 he was an inaugural scholar in AI ethics and Health at the University of Toronto's Joint Centre for Bioethics. His research interests broadly include patient and public involvement in science, technology, and innovation; and ethics and values in the design of emerging digital technologies. Twitter: @JosephDonia

INFO: 

This event is part of the Café Bioethics and Canadian Bioethics Society JUNE PARTNERSHIP Series! CB and CBS are collaborating for the first time in an effort to fill the summer conference gap created by the COVID-19 pandemic. We hope to provide students a platform to share their bioethics  research in an accessible manner, all while creating connections and building their network. Be sure to join us for the series!