This chapter examines the epistemology of algorithms, framing the discussion as a question of epistemic justification. Current approaches emphasize algorithmic transparency, which involves elucidating internal mechanisms—such as functions and variables—and demonstrating how (or that) these compute outputs. Thus, the mode of justification through transparency is contingent on what can be shown about the algorithm and, in this sense, is internal to the algorithm. In contrast, I propose an externalist epistemology of algorithms called computational reliabilism (CR). While I have previously developed CR in the context of computer simulations (Durán, Explaining simulated phenomena: A defense of the epistemic power of computer simulations, 2013; Durán, Computer simulations in science and engineering. Concepts - practices - perspectives. Springer, 2018; Durán, Formanek, Minds and Machines 28(4), 645–666, 2018), this chapter extends the framework to a broader range of algorithms used across scientific disciplines, particularly in machine learning and deep neural networks. At its core, CR posits that an algorithm’s output is justified if it is generated by a reliable algorithm, where reliability is determined by reliability indicators. These indicators arise from formal methods, algorithmic metrics, expert competencies, research cultures, and other scientific practices. The chapter’s primary objectives are to delineate the foundations of CR, explain its operational mechanisms, and outline its potential as an externalist epistemology of algorithms.

Achieving trustworthy AI is increasingly considered an essential desideratum to integrate AI systems into sensitive societal fields, such as criminal justice, finance, medicine, and healthcare, among others. For this reason, it is important to spell out clearly its characteristics, merits, and shortcomings. This article is the first survey in the specialized literature that maps out the philosophical landscape surrounding trust and trustworthiness in AI. To achieve our goals, we proceed as follows. We start by discussing philosophical positions on trust and trustworthiness, focusing on interpersonal accounts of trust. This allows us to explain why trust, in its most general terms, is to be understood as reliance plus some “extra factor”. We then turn to the first part of the definition provided, i.e., reliance, and analyze two opposing approaches to establishing AI systems’ reliability. On the one hand, we consider transparency and, on the other, computational reliabilism. Subsequently, we focus on debates revolving around the “extra factor”. To this end, we consider viewpoints that most actively resist the possibility and desirability of trusting AI systems before turning to the analysis of the most prominent advocates of it. Finally, we take up the main conclusions of the previous sections and briefly point at issues that remain open and need further attention.

In a reliabilist epistemology of algorithms, a high frequency of accurate output representations is indicative of the algorithm’s reliability. Recently, Humphreys challenged this assumption, arguing that reliability depends not only on frequency but also on the quality of outputs. Specifically, he contends that radical and egregious misrepresentations have a distinct epistemic impact on our assessment of an algorithm’s reliability, regardless of the frequency of their occurrence. He terms these statistically insignificant but serious errors (SIS-Errors) and maintains that their occurrence warrants revoking our epistemic attitude towards the algorithm’s reliability. This article seeks to defend reliabilist epistemologies of algorithms against the challenge posed by SIS-Errors. To this end, I draw upon computational reliabilism as a foundational framework and articulate epistemological conditions designed to prevent SIS-Errors and thus preserve algorithmic reliability.

Scientific explanations form an integral part of robust scientific practice, offering avenues for understanding, advancing scientific knowledge, and informed decision-making. This chapter delves into explanatory AI (artificial intelligence) (XAI), emphasizing the role of scientific explanations in the application of machine learning in the medical and healthcare domains. After offering a brief overview of the merits of employing AI in the fields of medicine and healthcare, as well as highlighting the value of XAI for physicians, patients, and institutions, this chapter establishes a core distinction between how-explanations and why-explanations. This distinction is essential for obtaining a deeper understanding of the complexities surrounding the XAI debate. This chapter also highlights that while how-explanations have received significant attention within more technically inclined literature, including philosophical discourse, why-explanations have somewhat lagged in exploration. This is unfortunate, since explaining why an algorithm suggests a given treatment recommends a give drug dosage or determines that a mole is malignant and carries enormous epistemic and moral value. Despite this, this chapter concludes optimistically, highlighting the increasing prominence of why-explanations in the ongoing discussions on XAI. Overall, this chapter illuminates the dynamic landscape of AI in the fields of medicine and healthcare, underscoring the significance and indispensability of XAI, alongside the challenges that lie ahead.

In this paper, we discuss epistemic and ethical concerns brought about by machine learning (ML) systems implemented in medicine. We begin by fleshing out the logic underlying a common approach in the specialized literature (which we call the informativeness account). We maintain that the informativeness account limits its analysis to the impact of epistemological issues on ethical concerns without assessing the bearings that ethical features have on the epistemological evaluation of ML systems. We argue that according to this methodological approach, epistemological issues are instrumental to and autonomous of ethical considerations. This means that the informativeness account considers epistemological evaluation uninfluenced and unregulated by an ethical counterpart. Using an example that does not square well into the informativeness account, we argue for ethical assessments that have a substantial influence on the epistemological assessment of ML and that such influence should not be understood as merely informative but rather regulatory. Drawing on the case analyzed, we claim that within the theoretical framework of the informativeness approach, forms of epistemic injustice—especially epistemic objectification—remain unaddressed. Our analysis should motivate further research investigating the regulatory role that ethical elements play in the epistemology of ML.

This chapter introduces philosophical debates on simulation models and their implementation as computer simulations. Two main views are identified, and a third is outlined. According to the first view, simulation models are taken to be mathematical models simpliciter and computer simulations are simply numerical solutions to those mathematical models. In the second view, simulation models are regarded as a new type of mathematical model with their own methodology, and computer simulations are viewed as a distinctive kind of scientific methodology. The third view is exploratory and sees simulation models as distant from mathematical modeling and as units of analysis in their own right. The chapter concludes with a discussion on epistemic opacity as an essential philosophical problem for computer simulations.

The social aspects of causality in medicine and healthcare have been emphasized in recent debates in the philosophy of science as crucial factors that need to be considered to enable, among others, appropriate interventions in public health. Therefore, it seems central to recognize the bearing of social causes (broadly understood, e.g., social inequalities and socio-economic status) in bringing about certain concrete pathologies. Being aware of the relevance of social causes in medicine and healthcare is particularly important in the face of the role that artificial intelligence-based systems (such as machine learning algorithms) are increasingly playing in these high-stakes fields. In fact, these systems bear the dangerous potential of concealing relevant social causes. This is highly problematic not only because it reinforces issues of distributive injustice but also because it can pave the way for issues of epistemic injustice. The central aim of this chapter is to make a first effort to point out possible connections between the importance of recognizing social causes in medicine and healthcare and forms of epistemic injustice.

Explanatory Artificial Intelligence (XAI) is receiving considerable attention from philosophers of science. A prevalent strategy is to integrate machine learning (ML) algorithms into existing accounts of scientific explanation. This chapter delves into causal approaches, specifically examining two recent interpretations of XAI: “the explanation game” by Watson and Floridi, and Buijsman's causal counterfactual approach. Both aim to integrate ML into Woodward's causal-manipulationist framework. Both have the goal of answering how an algorithm generates a given output. While making significant progress within the debate on XAI, both approaches fall short in key considerations about scientific explanation, particularly in addressing why-seeking questions. This chapter explores the conditions under which said questions can be answered. Two standard issues in the literature on philosophy of scientific explanation are addressed: identifying explanatory relations of dependence and (an alternative to) the factivity condition.

Machine learning techniques are driving — or soon will be driving — much of scientific research and discovery. Can they function as models similar to more traditional modeling techniques in scientific contexts? Or might they replace models altogether if they deliver sufficient predictive accuracy? These questions cut across at least two types of applications of machine learning models. First, machine learning models are used to study the brain, where neural networks might represent aspects of neural activity. The principal question here is: Can the uses of neural networks provide scientific explanations and models for neuroscience? Second, machine learning models are applied in science more broadly, where representational links are less clear. What are the epistemic implications of machine learning in those areas? Can they replace more traditional scientific models of phenomena, or is mere predictive accuracy sufficient? These two strands are finally brought together to create an overview of the (epistemic) role machine learning can play in scientific modeling.

In this article, I focus on the role of computer simulations as exploratory strategies. I begin by establishing the non-theory-driven nature of simulations. This refers to their ability to characterize phenomena without relying on a predefined conceptual framework that is provided by an implemented mathematical model. Drawing on Steinle’s notion of exploratory experimentation and Gelfert’s work on exploratory models, I present three exploratory strategies for computer simulations: (1) starting points and continuation of scientific inquiry, (2) varying the parameters, and (3) scientific prototyping.

The philosophical study of computer simulations has been largely subordinated to the analysis of sets of equations and their implementation on the computer. What has received less attention, however, is whether simulation models can be taken as units of analysis in their own right. Here I present my own experimental work investigating this issue. This article explores the capacity of programming languages to represent target systems and submits that, in a number of cases, the representation of simulation models differs in non-trivial ways from sets of equations. If my claim is correct, then a few important methodological and epistemological concerns emerge that need our attention. This article finishes by briefly addressing some implications for the philosophy of computer simulation.

AI is believed to have the potential to radically change modern medicine. Medical AI systems are developed to improve diagnosis, prediction, and treatment of a wide array of medical conditions. It is assumed to enable more accurate and efficient ways to diagnose diseases and “to restore the precious and time-honored connection and trust – the human touch – between patients and doctors“ (Topol, 2019, p. 18), by enabling health care professionals to spend more time with their patients. Sophisticated self-learning AI systems that do not follow predetermined decision rules – often referred to as black-boxes (Esteva et al 2019; Shortliffe et al 2018) – have spawned philosophical debate: the black-box nature of AI systems is believed to be a major ethical challenge for the use of these systems in medicine and it remains disputed whether explainability is philosophically and computationally possible. This special issue focuses on the ethics and epistemology of explainability in medical AI broadly construed.

The use of black box algorithms in medicine has raised scholarly concerns due to their opaqueness and lack of trustworthiness. Concerns about potential bias, accountability and responsibility, patient autonomy and compromised trust transpire with black box algorithms. These worries connect epistemic concerns with normative issues. In this paper, we outline that black box algorithms are less problematic for epistemic reasons than many scholars seem to believe. By outlining that more transparency in algorithms is not always necessary, and by explaining that computational processes are indeed methodologically opaque to humans, we argue that the reliability of algorithms provides reasons for trusting the outcomes of medical artificial intelligence (AI). To this end, we explain how computational reliabilism, which does not require transparency and supports the reliability of algorithms, justifies the belief that results of medical AI are to be trusted. We also argue that several ethical concerns remain with black box algorithms, even when the results are trustworthy. Having justified knowledge from reliable indicators is, therefore, necessary but not sufficient for normatively justifying physicians to act. This means that deliberation about the results of reliable algorithms is required to find out what is a desirable action. Thus understood, we argue that such challenges should not dismiss the use of black box algorithms altogether but should inform the way in which these algorithms are designed and implemented. When physicians are trained to acquire the necessary skills and expertise, and collaborate with medical informatics and data scientists, black box algorithms can contribute to improving medical care.

Explanatory AI (XAI) is on the rise, gaining enormous traction with the computational community, policymakers, and philosophers alike. This article contributes to this debate by first distinguishing scientific XAI (sXAI) from other forms of XAI. It further advances the structure for bona fide sXAI, while remaining neutral regarding preferences for theories of explanations. Three core components are under study, namely, i) the structure for bona fide sXAI, consisting in elucidating the explanans, the explanandum, and the explanatory relation for sXAI: ii) the pragmatics of explanation, which includes a discussion of the role of multi-agents receiving an explanation and the context within which the explanation is given; and iii) a discussion on Meaningful Human Explanation, an umbrella concept for different metrics required for measuring the explanatory power of explanations and the involvement of human agents in sXAI. The kind of AI systems of interest in this article are those utilized in medicine and the healthcare system. The article also critically addresses current philosophical and computational approaches to XAI. Amongst the main objections, it argues that there has been a long-standing interpretation of classifications as explanation, when these should be kept separate.

Medical AI is increasingly being developed and tested to improve medical diagnosis, prediction and treatment of a wide array of medical conditions. Despite worries about the explainability and accuracy of such medical AI systems, it is reasonable to assume that they will be increasingly implemented in medical practice. Current ethical debates focus mainly on design requirements and suggest embedding certain values such as transparency, fairness, and explainability in the design of medical AI systems. Aside from concerns about their design, medical AI systems also raise questions with regard to physicians' responsibilities once these technologies are being implemented and used. How do physicians’ responsibilities change with the implementation of medical AI? Which set of competencies do physicians have to learn to responsibly interact with medical AI? In the present article, we will introduce the notion of forward-looking responsibility and enumerate through this conceptual lens a number of competencies and duties that physicians ought to employ to responsibly utilize medical AI in practice. Those include amongst others understanding the range of reasonable outputs, being aware of own experience and skill decline, and monitoring potential accuracy decline of the AI systems.

Many philosophical accounts of scientific models fail to distinguish between a simulation model and other forms of models. This failure is unfortunate because there are important differences pertaining to their methodology and epistemology that favor their philosophical understanding. The core claim presented here is that simulation models are rich and complex units of analysis in their own right, that they depart from known forms of scientific models in significant ways, and that a proper understanding of the type of model simulations are fundamental for their philosophical assessment. I argue that simulation models can be distinguished from other forms of models by the many algorithmic structures, representation relations, and new semantic connections involved in their architecture. In this article, I reconstruct a general architecture for a simulation model, one that faithfully captures the complexities involved in most scientific research with computer simulations. Furthermore, I submit that a new methodology capable of conforming such architecture into a fully functional, computationally tractable computer simulation must be in place. I discuss this methodology—what I call recasting—and argue for its philosophical novelty. If these efforts are heading towards the right interpretation of simulation models, then one can show that computer simulations shed new light on the philosophy of science. To illustrate the potential of my interpretation of simulation models, I briefly discuss simulation-based explanations as a novel approach to questions about scientific explanation.

When one wants to use citizen input to inform policy, what should the standards of informedness on the part of the citizens be? While there are moral reasons to allow every citizen to participate and have a voice on every issue, regardless of education and involvement, designers of participatory assessments have to make decisions about how to structure deliberations as well as how much background information and deliberation time to provide to participants. After assessing different frameworks for the relationship between science and society, we use Philip Kitcher's framework of Well-Ordered Science to propose an epistemic standard on how citizen deliberations should be structured. We explore what potential standards follow from this epistemic framework focusing on significance versus scientific and engineering expertise. We argue that citizens should be tutored on the historical context of why scientific questions became significant and deemed scientifically and socially valuable, and if citizens report that they are capable of weighing in on an issue then they should be able to do so. We explore what this standard can mean by looking at actual citizen deliberations tied to the 2014 NASA ECAST Asteroid Initiative Citizen forums. We code different vignettes of citizens debating alternative approaches for Mars exploration based upon what level of information seemed to be sufficient for them to feel comfortable in making a policy position. The analysis provides recommendations on how to design and assess future citizen assessments grounded in properly conveying the historical value context surrounding a scientific issue and trusting citizens to seek out sufficient information to deliberate.

A plausible reading of the FAIR principles (Wilkinson et al. 2016) is that they were introduced with the intention to eliminate, or at least reduce, Dark Data (Dark Data being data which is non-reusable). Arguably, they derive their worth from how successful they are in this endeavour (even if only in principle). In this paper, we highlight a class of data which we call Lost Data. Lost Data is data, knowledge of whose existence has been lost to relevant interested parties. Lost Data poses a problem for the FAIR principles in so far as it is demonstrable that Lost Data both i) comply with the standards set out in the FAIR principles, and ii) nonetheless qualify as dark, by its existence being unknown to those relevant interested parties, and it’s being non-reusable, as a result. If the FAIR principles should indeed be rightly understood as an attempt to eliminate Dark Data, as we suggest, Lost Data thus poses a significant problem. Lost Data serves as a demonstration that the FAIR principles fail in this regard, and thus undermines the motivation for them. Lost Data, then, highlights the need for the augmentation of the FAIR principles to plug this identifiable “gap” in which Lost Data sits. In this paper, along with highlighting this problem, we suggest several possible changes that might be made to the FAIR principles that will help assuage it.