There is a current debate about if, and in what sense, machine learning systems used in the medical context need to be explainable. Those arguing in favor contend these systems require post hoc explanations for each individual decision to increase trust and ensure accurate diagnoses. Those arguing against suggest the high accuracy and reliability of the systems is sufficient for providing epistemic justified beliefs without the need for explaining each individual decision. But, as we show, both solutions have limitations—and it (...) is unclear either address the epistemic worries of the medical professionals using these systems. We argue these systems do require an explanation, but an institutional explanation. These types of explanations provide the reasons why the medical professional should rely on the system in practice—that is, they focus on trying to address the epistemic concerns of those using the system in specific contexts and specific occasions. But ensuring that these institutional explanations are fit for purpose means ensuring the institutions designing and deploying these systems are transparent about the assumptions baked into the system. This requires coordination with experts and end-users concerning how it will function in the field, the metrics used to evaluate its accuracy, and the procedures for auditing the system to prevent biases and failures from going unaddressed. We contend this broader explanation is necessary for either post hoc explanations or accuracy scores to be epistemically meaningful to the medical professional, making it possible for them to rely on these systems as effective and useful tools in their practices. (shrink)

Recent scholarship in philosophy of science and technology has shown that scientific and technological decision making are laden with values, including values of a social, political, and/or ethical character. This paper examines the role of value judgments in the design of machine-learning systems generally and in recidivism-prediction algorithms specifically. Drawing on work on inductive and epistemic risk, the paper argues that ML systems are value laden in ways similar to human decision making, because the development and design of ML systems (...) requires human decisions that involve tradeoffs that reflect values. In many cases, these decisions have significant—and, in some cases, disparate—downstream impacts on human lives. After examining an influential court decision regarding the use of proprietary recidivism-prediction algorithms in criminal sentencing, Wisconsin v. Loomis, the paper provides three recommendations for the use of ML in penal systems. (shrink)

Normative issues associated with the design and implementation of population-based lung cancer screening policies are underexamined. This study was an exposition of the ethical justification for screening and potential ethical issues and their solutions in Canadian jurisdictions. A qualitative description study was conducted. Key informants, defined as policymakers, scientists and clinicians who develop and implement lung cancer screening policies in Canada, were purposively sampled and interviewed using a semi-structured guide informed by population-based disease screening principles and ethical issues in cancer (...) screening. Interview data were analyzed using qualitative content analysis. Fifteen key informants from seven provinces were interviewed. Virtually all justified screening by beneficence, describing that population benefits outweigh individual harms if high-risk people are screened in organized programs according to disease screening principles. Equity of screening access, stigma and lung cancer primary prevention were other ethical issues identified. Key informants prioritized beneficence over concerns for group-level justice issues when making decisions about whether to implement screening policies. This prioritization, though slight, may impede the implementation of screening policies in a way that effectively addresses justice issues, a goal likely to require justice theory and critical interpretation of disease screening principles. (shrink)

One prominent line of support for nudging in screening programs is the claim that nudging can help ‘bad choosers’ — that is, it can help some patients make choices more in line with their own values and preferences. In this article, I argue that due to the presence of epistemic risk in many screening programs, the argument that nudging can help ‘bad choosers’ should be revised or rejected. Expanding on the work of Biddle, J. B. 2020. Epistemic risks in cancer (...) screening: Implications for ethics and policy. Studies in History and Philosophy of Science Part C: Studies in History and Philosophy of Biological and Biomedical Sciences 79: 101200.), I argue that epistemic risk undermines the argument that nudging can help to promote patient autonomy in the context of screening. Specifically, I argue that epistemic risk results in the inclusion of non-patient values and preferences in the screening process, which challenges the claim that nudging can help patients make choices more in line with their own values and preferences. I present four reasons to think epistemic risk undermines the argument in this way: (1) conflicting values; (2) lack of transparency; (3) limited autonomy in opting out; (4) unjustified manipulation. The presence of epistemic risk in screening programs means that nudging may not always be an effective means of promoting patient autonomy and informed consent. As such, epistemic risk poses significant challenges to at least one ethical justification of nudging in screening programs, and raises further questions about the role of nudging in promoting patient decision-making. (shrink)