Harisan Unais Nasir

When healthcare resources are scarce, they ought to be distributed fairly across society. For some theories of distribution, an assessment of individual health risk is required for a fair distribution of both healthcare resources and burdens. Despite this requirement, prevailing theories underappreciate the cost of information on health risk. Information relevant to health risk will be ignored when these costs are too high, leading to a form of discrimination that I term “information discrimination”. This leads to a potential source of unfairness for any health system that aims to distribute the burdens of healthcare across individuals according to their expected health states. Individuals whose risks are cheap to identify will have their risks personalized whereas individuals whose risks are expensive to identify will have their risks spread across society or the collective. This paper discusses the implications of this unfairness on two types of distributive theories in healthcare: responsibility‐sensitive and responsibility‐insensitive.

A significant and important ethical tension in resource allocation and public health ethics is between utility and equity. We explore this tension between utility and equity in the context of health AI through an examination of a diagnostic AI screening tool for diabetic retinopathy developed by a team of researchers at Duke-NUS in Singapore. While this tool was found to be effective, it was not equally effective across every ethnic group in Singapore, being less effective for the minority Malay population than for the Chinese majority. We discuss the problematic normative nature of bias in health AI and explore the ways in which bias can interact with various forms of social inequalities. From there, we examine the specifics of the diabetic retinopathy case and weigh up specific trade-offs between utility and equity. Ultimately, we conclude that it is ethically permissible to prioritise utility over equity where certain criteria hold. Given that any medical AI is more likely than not to have lingering bias due to bias in the training data that may reflect other social inequalities, we argue that it is permissible to implement an AI tool with residual bias where: (1) its introduction reduces the influence of biases (even if overall inequality is worsened), and/or (2) where the utility gained is significant enough and shared across groups (even if unevenly).

We live in the Information Age. Advances over the past 50 years in computing technology have enabled ever-expanding capacity to generate, store, transfer, process and analyse information about people, societies, products, services, the environment—nearly every aspect of the world. In parallel, concerns over how such data is being used have emerged, focusing especially on issues of privacy and confidentiality. Yet as technological capabilities continue to expand, the debate over ethical uses of data inevitably has evolved as well. Whereas we used to just talk about data, now there is great concern over Big Data. This chapter explores what, exactly, Big Data is; the ethical ramifications of its emergence, particularly for consent and anonymization; and what sorts of policy solutions are being proposed to address these emerging challenges. We will not defend one particular approach, but hope this chapter can be a useful reference for those who are contemplating difficult questions surrounding responsible management of Big Data.