This article argues against the claim that democracy is a necessary condition of political legitimacy. Instead, I propose a weaker set of conditions. First, I explain the case for the necessity of democracy. This is that only democracy can address the ‘egalitarian challenge’, i.e. ‘if we are all equal, why should only some of us wield political power?’. I show that if democracy really is a necessary condition of political legitimacy, then (what I label) the problems of domestic justice and (...) of international legitimacy become intractable. I then argue that the egalitarian challenge is addressed where the requirements of (1) horizontal equality, (2) acceptable vertical inequality, and (3) publicity, are met and where (4) citizens have some institutionalized opportunity for a voice in decisions. I show that these conditions can be realized in non-democratic form and conclude by explaining how the four conditions can be employed to make the problems of domestic justice and of international legitimacy more tractable. Overall, my ambitions are limited. I do not offer an all-things-considered case against democracy but I do show that (some) forms of non-democratic government are permissible. (shrink)

In the past few years, machine learning (ML) tools have been implemented with success in the medical context. However, several practitioners have raised concerns about the lack of transparency—at the algorithmic level—of many of these tools; and solutions from the field of explainable AI (XAI) have been seen as a way to open the ‘black box’ and make the tools more trustworthy. Recently, Alex London has argued that in the medical context we do not need machine learning tools (...) to be interpretable at the algorithmic level to make them trustworthy, as long as they meet some strict empirical desiderata. In this paper, we analyse and develop London’s position. In particular, we make two claims. First, we claim that London’s solution to the problem of trust can potentially address another problem, which is how to evaluate the reliability of ML tools in medicine for regulatory purposes. Second, we claim that to deal with this problem, we need to develop London’s views by shifting the focus from the opacity of algorithmic details to the opacity of the way in which ML tools are trained and built. We claim that to regulate AI tools and evaluate their reliability, agencies need an explanation of how ML tools have been built, which requires documenting and justifying the technical choices that practitioners have made in designing such tools. This is because different algorithmic designs may lead to different outcomes, and to the realization of different purposes. However, given that technical choices underlying algorithmic design are shaped by value-laden considerations, opening the black box of the design process means also making transparent and motivating (technical and ethical) values and preferences behind such choices. Using tools from philosophy of technology and philosophy of science, we elaborate a framework showing how an explanation of the training processes of ML tools in medicine should look like. (shrink)

Objective: To examine the role of explainability in machine learning for healthcare (MLHC), and its necessity and significance with respect to effective and ethical MLHC application. Study Design and Setting: This commentary engages with the growing and dynamic corpus of literature on the use of MLHC and artificial intelligence (AI) in medicine, which provide the context for a focused narrative review of arguments presented in favour of and opposition to explainability in MLHC. Results: We find that concerns (...) regarding explainability are not limited to MLHC, but rather extend to numerous well-validated treatment interventions as well as to human clinical judgment itself. We examine the role of evidence-based medicine in evaluating unexplainable treatments and technologies, and highlight the analogy between the concept of explainability in MLHC and the related concept of mechanistic reasoning in evidence-based medicine. Conclusion: Ultimately, we conclude that the value of explainability in MLHC is not intrinsic, but is instead instrumental to achieving greater imperatives such as performance and trust. We caution against the uncompromising pursuit of explainability, and advocate instead for the development of robust empirical methods to successfully evaluate increasingly inexplicable algorithmic systems. (shrink)

The operations of deep networks are widely acknowledged to be inscrutable. The growing field of Explainable AI has emerged in direct response to this problem. However, owing to the nature of the opacity in question, XAI has been forced to prioritise interpretability at the expense of completeness, and even realism, so that its explanations are frequently interpretable without being underpinned by more comprehensive explanations faithful to the way a network computes its predictions. While this has been taken to be a (...) shortcoming of the field of XAI, I argue that it is broadly the right approach to the problem. (shrink)

Deep learning and other black-box models are becoming more and more popular today. Despite their high performance, they may not be accepted ethically or legally because of their lack of explainability. This paper presents the increasing number of legal requirements on machine learning model interpretability and explainability in the context of private and public decision making. It then explains how those legal requirements can be implemented into machine-learning models and concludes with a call for more inter-disciplinary (...) research on explainability. (shrink)

Machine learning (ML) models recently led to major breakthroughs in predictive tasks in the natural sciences. Yet their benefits for the social sciences are less evident, as even high-profile studies on the prediction of life trajectories have shown to be largely unsuccessful – at least when measured in traditional criteria of scientific success. This paper tries to shed light on this remarkable performance gap. Comparing two social science case studies to a paradigm example from the natural sciences, we argue (...) that, in addition to explanation, prediction is an important goal of social science – and we identify constraints that impede pure ML prediction from being successful in that field. As a remedy, we outline elements of an integrative modelling approach that combines explanatory models and predictive ML models. (shrink)

The machine conception of the organism (MCO) is one of the most pervasive notions in modern biology. However, it has not yet received much attention by philosophers of biology. The MCO has its origins in Cartesian natural philosophy, and it is based on the metaphorical redescription of the organism as a machine. In this paper I argue that although organisms and machines resemble each other in some basic respects, they are actually very different kinds of systems. I submit (...) that the most significant difference between organisms and machines is that the former are intrinsically purposive whereas the latter are extrinsically purposive. Using this distinction as a starting point, I discuss a wide range of dissimilarities between organisms and machines that collectively lay bare the inadequacy of the MCO as a general theory of living systems. To account for the MCO’s prevalence in biology, I distinguish between its theoretical, heuristic, and rhetorical functions. I explain why the MCO is valuable when it is employed heuristically but not theoretically, and finally I illustrate the serious problems that arise from the rhetorical appeal to the MCO. (shrink)

Machine learning systems are making considerable inroads in society owing to their ability to recognize and predict patterns. However, the decision-making logic of some widely used machine learning models, such as deep neural networks, is characterized by opacity, thereby rendering them exceedingly difficult for humans to understand and explain and, as a result, potentially risky to use. Considering the importance of addressing this opacity, this paper calls for research that studies empirically and theoretically how machine learning experts (...) and users seek to attain machine learning explainability. Focusing on automated trading, we take steps in this direction by analyzing a trading firm’s quest for explaining its deep neural network system’s actionable predictions. We demonstrate that this explainability effort involves a particular form of human–machine interaction that contains both anthropomorphic and technomorphic elements. We discuss this attempt to attain machine learning explainability in light of reflections on cross-species companionship and consider it an example of human–machine companionship. (shrink)

Analogies to machines are commonplace in the life sciences, especially in cellular and molecular biology — they shape conceptions of phenomena and expectations about how they are to be explained. This paper offers a framework for thinking about such analogies. The guiding idea is that machine-like systems are especially amenable to decompositional explanation, i.e., to analyses that tease apart underlying components and attend to their structural features and interrelations. I argue that for decomposition to succeed a system must exhibit (...) causal orderliness, which I explicate in terms of differentiation among parts and the significance of local relations. I also discuss what makes a model depict its target as machine-like, suggesting that a key issue is the degree of detail with respect to the target’s parts and their interrelations. (shrink)

Machine learning algorithms may radically improve our ability to diagnose and treat disease. For moral, legal, and scientific reasons, it is essential that doctors and patients be able to understand and explain the predictions of these models. Scalable, customisable, and ethical solutions can be achieved by working together with relevant stakeholders, including patients, data scientists, and policy makers.

Gödel's theorem seems to me to prove that Mechanism is false, that is, that minds cannot be explained as machines. So also has it seemed to many other people: almost every mathematical logician I have put the matter to has confessed to similar thoughts, but has felt reluctant to commit himself definitely until he could see the whole argument set out, with all objections fully stated and properly met.1 This I attempt to do.

Although machine learning has been successful in recent years and is increasingly being deployed in the sciences, enterprises or administrations, it has rarely been discussed in philosophy beyond the philosophy of mathematics and machine learning. The present contribution addresses the resulting lack of conceptual tools for an epistemological discussion of machine learning by conceiving of deep learning networks as ‘judging machines’ and using the Kantian analysis of judgments for specifying the type of judgment they are capable of. (...) At the center of the argument is the fact that the functionality of deep learning networks is established by training and cannot be explained and justified by reference to a predefined rule-based procedure. Instead, the computational process of a deep learning network is barely explainable and needs further justification, as is shown in reference to the current research literature. Thus, it requires a new form of justification, that is to be specified with the help of Kant’s epistemology. (shrink)

In this article, Lucas maintains the falseness of Mechanism - the attempt to explain minds as machines - by means of Incompleteness Theorem of Gödel. Gödel’s theorem shows that in any system consistent and adequate for simple arithmetic there are formulae which cannot be proved in the system but that human minds can recognize as true; Lucas points out in his turn that Gödel’s theorem applies to machines because a machine is the concrete instantiation of a formal system: therefore, (...) for every machine consistent and able of doing simple arithmetic, there is a formula that it can’t produce as true but that we can see to be true, and so human minds and machines have to be different. Lucas considers as well in this article some possible objections to his argument: for any Gödelian formula we could, for instance, construct a machine able to produce it or we could put the Gödelian formulae that we had proved as axioms of a further machine. However - as Lucas underlines - for every of such machines we could again formulate another Gödelian formula, the Gödelian formula of these machines, that they are not able to proof but that we can recognize as true. More general arguments, such as the possibility to escape Gödelian argument by suggesting that Gödel’s theorem applies to consistent systems while we could be inconsistent ones, are moreover refuted by Lucas by maintaining that our inconsistency corresponds to occasional malfunctioning of a machine and not to his normal inconsistency; indeed, a inconsistent machine is characterized by producing any statement, on the contrary human being are selective and not disposed to assert anything. (shrink)

While Nozick and his sympathizers assume there is a widespread anti-hedonist intuition to prefer reality to an experience machine, hedonists have marshalled empirical evidence that shows such an assumption to be unfounded. Results of several experience machine variants indicate there is no widespread anti-hedonist intuition. From these findings, hedonists claim Nozick's argument fails as an objection to hedonism. This article suggests the argument surrounding experience machines has been misconceived. Rather than eliciting intuitions about what is prudentially valuable, these (...) intuitive judgements are instead calculations about prudential pay-offs and trade-offs. This position can help explain the divergence of intuitions people have about experience machines. (shrink)

Modern explainable AI methods remain far from providing human-like answers to ‘why’ questions, let alone those that satisfactorily agree with human-level understanding. Instead, the results that such methods provide boil down to sets of causal attributions. Currently, the choice of accepted attributions rests largely, if not solely, on the explainee’s understanding of the quality of explanations. The paper argues that such decisions may be transferred from a human to an XAI agent, provided that its machine-learning algorithms perform genuinely abductive (...) inferences. The paper outlines the key predicament in the current inductive paradigm of ML and the associated XAI techniques, and sketches the desiderata for a truly participatory, second-generation XAI, which is endowed with abduction. (shrink)

The workshop “Machine Learning: Prediction Without Explanation?” brought together philosophers of science and scholars from various fields who study and employ Machine Learning (ML) techniques, in order to discuss the changing face of science in the light of ML's constantly growing use. One major focus of the workshop was on the impact of ML on the concept and value of scientific explanation. One may speculate whether ML’s increased use in science exemplifies a paradigmatic turn towards mere pattern recognition (...) and prediction and away from science’s traditional aim of explanation. In contrast, certain conceptions of explanation, such as statistical explanation, could turn out to fit well with the achievements of present-day ML and concede an explanatory value to these achievements after all. It is an open question how to explain ML successes themselves, and this question was in the focus of several talks in the workshop. Based on the topics raised, we will discuss the talks’ contents in more detail as organized into (i) practitioners’ perspectives, (ii) explanations from ML, (iii) explanations of ML, (iv) societal implications, and (v) global and historical perspectives. (shrink)

In, “The Turing Triage Test”, published in Ethics and Information Technology, I described a hypothetical scenario, modelled on the famous Turing Test for machine intelligence, which might serve as means of testing whether or not machines had achieved the moral standing of people. In this paper, I: (1) explain why the Turing Triage Test is of vital interest in the context of contemporary debates about the ethics of AI; (2) address some issues that complexify the application of this test; (...) and, (3) in doing so, defend a way of thinking about the question of the moral standing of intelligent machines, which takes the idea of “seriousness” seriously. This last objective is, in fact, my primary one and is motivated by the sense that, to date, much of the “philosophy” of AI has suffered from a profound failure to properly distinguish between things that we can say and things that we can really mean. (shrink)

Should we give moral standing to machines? In this paper, I explore the implications of a relational approach to moral standing for thinking about machines, in particular autonomous, intelligent robots. I show how my version of this approach, which focuses on moral relations and on the conditions of possibility of moral status ascription, provides a way to take critical distance from what I call the “standard” approach to thinking about moral status and moral standing, which is based on properties. It (...) does not only overcome epistemological problems with the standard approach, but can also explain how we think about, experience, and act towards machines—including the gap that sometimes occurs between reasoning and experience. I also articulate the non-Cartesian orientation of my “relational” research program and specify the way it contributes to a different paradigm in thinking about moral standing and moral knowledge. (shrink)

As machine learning has gradually entered into ever more sectors of public and private life, there has been a growing demand for algorithmic explainability. How can we make the predictions of complex statistical models more intelligible to end users? A subdiscipline of computer science known as interpretable machine learning (IML) has emerged to address this urgent question. Numerous influential methods have been proposed, from local linear approximations to rule lists and counterfactuals. In this article, I highlight three (...) conceptual challenges that are largely overlooked by authors in this area. I argue that the vast majority of IML algorithms are plagued by (1) ambiguity with respect to their true target; (2) a disregard for error rates and severe testing; and (3) an emphasis on product over process. Each point is developed at length, drawing on relevant debates in epistemology and philosophy of science. Examples and counterexamples from IML are considered, demonstrating how failure to acknowledge these problems can result in counterintuitive and potentially misleading explanations. Without greater care for the conceptual foundations of IML, future work in this area is doomed to repeat the same mistakes. (shrink)

This paper provides an analysis of the way in which two foundational principles of medical ethics–the trusted doctor and patient autonomy–can be undermined by the use of machine learning (ML) algorithms and addresses its legal significance. This paper can be a guide to both health care providers and other stakeholders about how to anticipate and in some cases mitigate ethical conflicts caused by the use of ML in healthcare. It can also be read as a road map as to (...) what needs to be done to achieve an acceptable level of explainability in an ML algorithm when it is used in a healthcare context. (shrink)

As scientists start to adopt machine learning (ML) as one research tool, the security of ML and the knowledge generated become a concern. In this paper, I explain how supervised ML can be improved with better data ontology, or the way we make categories and turn information into data. More specifically, we should design data ontology in such a way that is consistent with the knowledge that we have about the target phenomenon so that such ontology can help us (...) make the inductive leap. I do so by thinking through a thought experiment, Goodman’s New Riddle of Induction (Fact, fiction, and forecast, Harvard University Press, 1955). Goodman’s riddle helps flesh out three problems of induction: (1) the problem of equal goodies, that there are often too many equally good inductive results given the same data; (2) the problem of diverging performance, that these equally good results can give opposite predictions in the future; and (3) the problem of mediocrity, that when averaged across all equally possible datasets and tasks, no inductive algorithm outperforms any other. I show that all these three problems are manifested as real obstacles in ML practice, namely, the Rashomon effect (Breiman in Stat Sci 16(3):199–231, 2001), the problem of underspecification (D’Amour et al. in J Mach Learn Res, 2020, https://doi.org/10.48550/arXiv.2011.03395 ), and the No Free Lunch theorem (Wolpert in Neural Comput 8(7):1341–90, 1996, https://doi.org/10.1162/neco.1996.8.7.1341 ). Lastly, I argue that proper data ontology can help mitigate these problems and I demonstrate how using concrete examples from climate science. This research highlights the links between philosophers’ discussions of induction and implications in ML practice. (shrink)

Machine learning-based AI algorithms lack transparency. In this article, I offer an interpretation of AI’s explainability problem and highlight its ethical saliency. I try to make the case for the legal enforcement of a strong explainability requirement: human organizations which decide to automate decision-making should be legally obliged to demonstrate the capacity to explain and justify the algorithmic decisions that have an impact on the wellbeing, rights, and opportunities of those affected by the decisions. This legal duty (...) can be derived from the demands of Rawlsian public reason. In the second part of the paper, I try to show that the argument from the limitations of human cognition fails to get AI off the hook of public reason. Against a growing trend in AI ethics, my main argument is that the analogy between human minds and artificial neural networks fails because it suffers from an atomistic bias which makes it blind to the social and institutional dimension of human reasoning processes. I suggest that developing interpretive AI algorithms is not the only possible answer to the explainability problem; social and institutional answers are also available and in many cases more trustworthy than techno-scientific ones. (shrink)

The notion of computation has changed the world more than any previous expressions of knowledge. However, as know-how in its particular algorithmic embodiment, computation is closed to meaning. Therefore, computer-based data processing can only mimic life’s creative aspects, without being creative itself. AI’s current record of accomplishments shows that it automates tasks associated with intelligence, without being intelligent itself. Mistaking the abstract for the concrete has led to the religion of “everything is an output of computation”—even the humankind that conceived (...) the computer. The hypostatized role of computers explains the increased dependence on them. The convergence machine called deep learning is only the most recent form through which the deterministic theology of the machine claims more than what it actually is: extremely effective data processing. A proper understanding of complexity, as well as the need to distinguish between the reactive nature of the artificial and the anticipatory nature of the living are suggested as practical responses to the challenges posed by machine theology. (shrink)

In this paper, I reconstruct Robert Nozick's experience machine objection to hedonism about well-being. I then explain and briefly discuss the most important recent criticisms that have been made of it. Finally, I question the conventional wisdom that the experience machine, while it neatly disposes of hedonism, poses no problem for desire-based theories of well-being.

Machine learning and statistical methods have yielded impressive results in a wide variety of natural language processing tasks. These advances have generally been regarded as engineering achievements. In fact it is possible to argue that the success of machine learning methods is significant for our understanding of the cognitive basis of language acquisition and processing. Recent work in unsupervised grammar induction is particularly relevant to this issue. It suggests that knowledge of language can be achieved through general learning (...) procedures, and that a richly articulated language faculty is not required to explain its acquisition. (shrink)

Machine vision is one of the main applications of artificial intelligence. In China, the machine vision industry makes up more than a third of the national AI market, and technologies like face recognition, object tracking and automated driving play a central role in surveillance systems and social governance projects relying on the large-scale collection and processing of sensor data. Like other novel articulations of technology and society, machine vision is defined, developed and explained by different actors through (...) the work of imagination. In this article, we draw on the concept of sociotechnical imaginaries to understand how Chinese companies represent machine vision. Through a qualitative multimodal analysis of the corporate websites of leading industry players, we identify a cohesive sociotechnical imaginary of machine vision, and explain how four distinct visual registers contribute to its articulation. These four registers, which we call computational abstraction, human–machine coordination, smooth everyday, and dashboard realism, allow Chinese tech companies to articulate their global ambitions and competitiveness through narrow and opaque representations of machine vision technologies. (shrink)

The essays in this volume represent the first steps by philosophers and artificial intelligence researchers toward explaining why it is necessary to add an ...

We argue why interpretability should have primacy alongside empiricism for several reasons: first, if machine learning models are beginning to render some of the high-risk healthcare decisions instead of clinicians, these models pose a novel medicolegal and ethical frontier that is incompletely addressed by current methods of appraising medical interventions like pharmacological therapies; second, a number of judicial precedents underpinning medical liability and negligence are compromised when ‘autonomous’ ML recommendations are considered to be en par with human instruction in (...) specific contexts; third, explainable algorithms may be more amenable to the ascertainment and minimisation of biases, with repercussions for racial equity as well as scientific reproducibility and generalisability. We conclude with some reasons for the ineludible importance of interpretability, such as the establishment of trust, in overcoming perhaps the most difficult challenge ML will face in a high-stakes environment like healthcare: professional and public acceptance. (shrink)

This study investigates the performance of 21 monolingual and 56 bilingual children aged 5;6-9;0 on German-LITMUS-sentence-repetition (SRT; Hamann et al., 2013) and nonword-repetition-tasks (NWRT; Grimm et al., 2014), which were constructed according to the LITMUS-principles (Language Impairment Testing in Multilingual Settings; Armon-Lotem et al., 2015). Both tasks incorporate complex structures shown to be cross-linguistically challenging for children with Specific Language Impairment (SLI) and aim at minimizing bias against bilingual children while still being indicative of the presence of language impairment across (...) language combinations. Given the great variability in bilingual language exposure and the potential effect of language experience on language performance in bilingual children, we examined whether background variables related to bilingualism, particularly, the degree language dominance as measured by relative amount of use and exposure, could compromise the diagnostic accuracy of the German-LITMUS-SRT and NWRT. We further investigated whether a combination of the two tasks provides better diagnostic accurracy. To address this, we used an unsupervised machine learning algorithm, the Partitioning-Around-Medoids (PAM), for deriving a clinical category for the children as ±language-impaired based on performance scores on SRT and/or NWRT while withholding information about their clinical status based on standardized assessment in L1/L2. Subsequently, we calculated diagnostic accuracy and investigated which background variables (age of onset, length of exposure, degree of language-dominance, socio-economic-status, and risk factors for SLI) best explained clinical-group-membership yielded from the PAM-analysis based on NWRT and/or SRT scores. Results show that although language-dominance influences the performance of bilingual typically-developing children, especially in the SRT, the diagnostic accuracy of the tools is not compromised by language dominance: while risk factors for SLI were significant predictors for clinical group membership in all models, language dominance did not contribute at all to explaining clinical cluster membership as typically-developing or SLI based on any of the combinations of the SRT and NWRT variables. Additionally, results confirm that a combination of SRT scored by correct target structure and the structurally more complex language-dependent part of the NWRT yields better diagnostic accuracy than single measures and is only sensitive for risk factors for SLI and not to dominance levels or SES. (shrink)

According to the conventional wisdom, Turing said that computing machines can be intelligent. I don't believe it. I think that what Turing really said was that computing machines –- computers limited to computing –- can only fake intelligence. If we want computers to become genuinelyintelligent, we will have to give them enough “initiative” to do more than compute. In this paper, I want to try to develop this idea. I want to explain how giving computers more ``initiative'' can allow them (...) to do more than compute. And I want to say why I believe that they will have to go beyond computation before they can become genuinely intelligent. (shrink)

As we increasingly delegate decision-making to algorithms, whether directly or indirectly, important questions emerge in circumstances where those decisions have direct consequences for individual rights and personal opportunities, as well as for the collective good. A key problem for policymakers is that the social implications of these new methods can only be grasped if there is an adequate comprehension of their general technical underpinnings. The discussion here focuses primarily on the case of enforcement decisions in the criminal justice system, but (...) draws on similar situations emerging from other algorithms utilised in controlling access to opportunities, to explain how machine learning works and, as a result, how decisions are made by modern intelligent algorithms or 'classifiers'. It examines the key aspects of the performance of classifiers, including how classifiers learn, the fact that they operate on the basis of correlation rather than causation, and that the term 'bias' in machine learning has a different meaning to common usage. An example of a real world 'classifier', the Harm Assessment Risk Tool (HART), is examined, through identification of its technical features: the classification method, the training data and the test data, the features and the labels, validation and performance measures. Four normative benchmarks are then considered by reference to HART: (a) prediction accuracy (b) fairness and equality before the law (c) transparency and accountability (d) informational privacy and freedom of expression, in order to demonstrate how its technical features have important normative dimensions that bear directly on the extent to which the system can be regarded as a viable and legitimate support for, or even alternative to, existing human decision-makers. (shrink)

Rapid advancements in machine learning techniques allow mass surveillance to be applied on larger scales and utilize more and more personal data. These developments demand reconsideration of the privacy-security dilemma, which describes the tradeoffs between national security interests and individual privacy concerns. By investigating mass surveillance techniques that use bulk data collection and machine learning algorithms, we show why these methods are unlikely to pinpoint terrorists in order to prevent attacks. The diverse characteristics of terrorist attacks—especially when considering (...) lone-wolf terrorism—lead to irregular and isolated footprints. The irregularity of data affects the accuracy of machine learning algorithms and the mass surveillance that depends on them which can be explained by three kinds of known problems encountered in machine learning theory: class imbalance, the curse of dimensionality, and spurious correlations. Proponents of mass surveillance often invoke the distinction between collecting data and metadata, in which the latter is understood as a lesser breach of privacy. Their arguments commonly overlook the ambiguity in the definitions of data and metadata and ignore the ability of machine learning techniques to infer the former from the latter. Given the sparsity of datasets used for machine learning in counterterrorism and the privacy risks attendant with bulk data collection, policymakers and other relevant stakeholders should critically re-evaluate the likelihood of success of the algorithms and the collection of data on which they depend. (shrink)

The focus of this article is a question that has been neglected in debates about digitalization: Could machines replace human scientists? To provide an intelligible answer to it, we need to answer a further question: What is it that makes (or constitutes) a scientist? I offer an answer to this question by proposing a new demarcation criterion for science which I call “the discoverability criterion”. I proceed as follows: (1) I explain why the target question of this article is important, (...) and (2) show that it leads to a variant of the demarcation problem of science. (3) By arguing that it is probably an essential feature of science that we can make scientific discoveries, I suggest a novel way of dealing with this problem by proposing a new demarcation criterion. Before introducing it, (4) I analyze an exemplary case of a real scientific discovery, and (5) argue that scientific discovery processes have a general underlying structure. (6) I introduce my discoverability criterion for science and present my master argument that helps us understand which criteria have to be fulfilled in order to decide whether machines can replace human scientists or not. (7) I conclude by answering the article’s target question and bringing forward a take-home message. (shrink)

This book provides a framework for thinking about foundational philosophical questions surrounding machine learning as an approach to artificial intelligence. Specifically, it links recent breakthroughs in deep learning to classical empiricist philosophy of mind. In recent assessments of deep learning's current capabilities and future potential, prominent scientists have cited historical figures from the perennial philosophical debate between nativism and empiricism, which primarily concerns the origins of abstract knowledge. These empiricists were generally faculty psychologists; that is, they argued that the (...) active engagement of general psychological faculties-such as perception, memory, imagination, attention, and empathy-enables rational agents to extract abstract knowledge from sensory experience. This book explains a number of recent attempts to model roles attributed to these faculties in deep neural network based artificial agents by appeal to the faculty psychology of philosophers such as Aristotle, Ibn Sina (Avicenna), John Locke David Hume, William James, and Sophie de Grouchy. It illustrates the utility of this interdisciplinary connection by showing how it can provide benefits to both philosophy and computer science: computer scientists can continue to mine the history of philosophy for ideas and aspirational targets to hit on the way to more robustly rational artificial agents, and philosophers can see how some of the historical empiricists' most ambitious speculations can be realized in specific computational systems. (shrink)

This paper elucidates the philosophical origins of the conception of plants as machines and analyses the contemporary technical and ethical consequences of that thinking. First, we explain the historical relationship between the explicit animal machine thesis of Descartes and the implicit plant machine thesis of today. Our hypothesis is that, although it is rarely discussed, the plant machine thesis remains influential. We define the philosophical criteria for both a moderate and radical interpretation of the thesis. Then, assessing (...) the compatibility of current botanical knowledge with both interpretations, we find that neither withstands scrutiny. We trace how biological and agricultural sciences have historically relied upon thinking of plants as machines and how they continue to do so today through rhetoric centred on breeding, biotechnology, and production. We discuss some of the most important legal and ethical consequences of obscuring the vitality of plants. Finally, we explore less reductive and destructive ways of thinking about, and using, plants. (shrink)

Machine learning is used more and more in scientific contexts, from the recent breakthroughs with AlphaFold2 in protein fold prediction to the use of ML in parametrization for large climate/astronomy models. Yet it is unclear whether we can obtain scientific explanations from such models. I argue that when machine learning is used to conduct causal inference we can give a new positive answer to this question. However, these ML models are purpose-built models and there are technical results showing (...) that standard machine learning models cannot be used for the same type of causal inference. Instead, there is a pathway to causal explanations from predictive ML models through new explainability techniques; specifically, new methods to extract structural equation models from such ML models. The extracted models are likely to suffer from issues though: they will often fail to account for confounders and colliders, as well as deliver simply incorrect causal graphs due to ML models tendency to violate physical laws such as the conservation of energy. In this case, extracted graphs are a starting point for new explanations, but predictive accuracy is no guarantee for good explanations. (shrink)

This paper establishes a connection between the fields of machine learning (ML) and philosophy concerning the phenomenon of behaving neutrally. It investigates a specific class of ML systems capable of delivering a neutral response to a given task, referred to as abstaining machine learning systems, that has not yet been studied from a philosophical perspective. The paper introduces and explains various abstaining machine learning systems, and categorizes them into distinct types. An examination is conducted on how abstention (...) in the different machine learning system types aligns with the epistemological counterpart of suspended judgment, addressing both the nature of suspension and its normative profile. Additionally, a philosophical analysis is suggested on the autonomy and explainability of the abstaining response. It is argued, specifically, that one of the distinguished types of abstaining systems is preferable as it aligns more closely with our criteria for suspended judgment. Moreover, it is better equipped to autonomously generate abstaining outputs and offer explanations for abstaining outputs when compared to the other type. (shrink)

Theories of action tend to require agents to have mental representations. A common trope in discussions of artificial intelligence (AI) is that they do not, and so cannot be agents. Properly understood there may be something to the requirement, but the trope is badly misguided. Here we provide an account of representation for AI that is sufficient to underwrite attributions to these systems of ownership, action, and responsibility. Existing accounts of mental representation tend to be too demanding and unparsimonious. We (...) offer instead a minimalist account of representation that ascribes only those features necessary for explaining action, trimming the “extra” features in existing accounts (e.g., representation as a “mental” phenomenon). Our account makes ‘representation’ whatever it is that, for example, the thermostat is doing with the thermometer. The thermostat is disposed to act as long as the thermometer is outside a given range of parameters.. Our account allows us to offer a new perspective on the ‘responsibility gap’, a problem raised by the actions of sophisticated machines: because nobody has enough control over the machine’s actions to be able to assume responsibility, conventional approaches to responsibility ascription are inappropriate. We argue that there is a distinction between finding responsible and holding responsible and, in order to resolve the responsibility gap, we must first clarify the conceptual terrain on which agent is in fact responsible. (shrink)

An analysis of how capitalism today produces subjectivity like any other “good,” and what would allow us to escape its hold. “Capital is a semiotic operator”: this assertion by Félix Guattari is at the heart of Maurizio Lazzarato's Signs and Machines, which asks us to leave behind the logocentrism that still informs so many critical theories. Lazzarato calls instead for a new theory capable of explaining how signs function in the economy, in power apparatuses, and in the production of subjectivity. (...) Moving beyond the dualism of signifier and signified, Signs and Machines shows how signs act as “sign-operators” that enter directly into material flows and into the functioning of machines. Money, the stock market, price differentials, algorithms, and scientific equations and formulas constitute semiotic “motors” that make capitalism's social and technical machines run, bypassing representation and consciousness to produce social subjections and semiotic enslavements. Lazzarato contrasts Deleuze and Guattari's complex semiotics with the political theories of Jacques Rancière, Antonio Negri and Michael Hardt, Paolo Virno, and Judith Butler, for whom language and the public space it opens still play a fundamental role. Lazzarato asks: What are the conditions necessary for political and existential rupture at a time when the production of subjectivity represents the primary and perhaps most important work of capitalism? What are the specific tools required to undo the industrial mass production of subjectivity undertaken by business and the state? What types of organization must we construct for a process of subjectivation that would allow us to escape the hold of social subjection and machinic enslavement? In addressing these questions, Signs and Machines takes on a task that is today more urgent than ever. (shrink)

In The Sleepwalkers and The Act of Creation Arthur Koestler provided pioneering studies of scientific discovery and artistic inspiration, the twin pinnacles of human achievement. The Ghost in the Machine looks at the dark side of the coin: our terrible urge to self-destruction... Could the human species be a gigantic evolutionary mistake? To answer that startling question Koestler examines how experts on evolution and psychology all too often write about people with an 'antiquated slot-machine model based on the (...) naively mechanistic world-view of the nineteenth century. His brilliant polemic helped to instigate a major revolution in the life sciences, yet its 'glimpses of an alternative world-view' form only the background to an even more challenging analysis of the human predicament. Perhaps, he suggests, we are a species in which ancient and recent brain structures - or reason and emotion - are not fully co-ordinated. Such in-built deficiencies may explain the paranoia, violence and insanity that are central strands of human history. And however disturbing we find such issues, Koestler contends, it is only when we face our limitations head-on that we can hope to find a remedy. (shrink)

Recent work on interpretability in machine learning and AI has focused on the building of simplified models that approximate the true criteria used to make decisions. These models are a useful pedagogical device for teaching trained professionals how to predict what decisions will be made by the complex system, and most importantly how the system might break. However, when considering any such model it’s important to remember Box’s maxim that "All models are wrong but some are useful." We focus (...) on the distinction between these models and explanations in philosophy and sociology. These models can be understood as a "do it yourself kit" for explanations, allowing a practitioner to directly answer "what if questions" or generate contrastive explanations without external assistance. Although a valuable ability, giving these models as explanations appears more difficult than necessary, and other forms of explanation may not have the same trade-offs. We contrast the different schools of thought on what makes an explanation, and suggest that machine learning might benefit from viewing the problem more broadly. (shrink)

After proposing the Turing Test, Alan Turing himself considered a number of objections to the idea that a machine might eventually pass it. One of the objections discussed by Turing was that no machine will ever pass the Turing Test because no machine will ever “have as much diversity of behaviour as a man”. He responded as follows: the “criticism that a machine cannot have much diversity of behaviour is just a way of saying that it (...) cannot have much storage capacity”. I shall argue that the objection cannot be dismissed so easily. The diversity exhibited by human behaviour is characterized by a kind of context-sensitive adaptive plasticity. Most of the time, human beings flexibly and fluently respond to what is relevant in a given situation. Moreover, ordinary human life involves an open-ended flow of shifting contexts to which our behaviour typically adapts in real time. For a machine to “have as much diversity of behaviour as a man” would be for that machine to keep its responses and behaviour relevant within such a flow. Merely giving a machine the capacity to store a huge amount of information and an enormous number of behaviour-generating rules will not achieve this goal. By drawing on arguments presented originally by Descartes, and by making contact with the frame problem in artificial intelligence, I shall argue that the distinctive context-sensitive adaptive plasticity of human behaviour explains why the Turing Test is such a stringent test for the presence of thought, and why it is much harder to pass than Turing himself may have realized. (shrink)

This article examines the debate among natural philosophers during the early modern period which concerned whether living beings could be understood as biological machines that did not require a distinct principle of life or soul to explain their complex functioning. It suggests that these innovations can be seen collectively as a gradual substitution of the categorial framework of Aristotle by one derived from the experimental and mathematical sciences. The traditional epistemic relationship between natural philosophy and metaphysics thereby began a long-term (...) reversal. This article also discusses the sources of the early modern machine concept, animal spirits, and the problem of intermediate principles. (shrink)

In this article I explore various facets of Nozick’s famous thought experiment involving the experience machine. Nozick’s original target is hedonism—the view that the only intrinsic prudential value is pleasure. But the argument, if successful, undermines any experientialist theory, i.e. any theory that limits intrinsic prudential value to mental states. I first highlight problems arising from the way Nozick sets up the thought experiment. He asks us to imagine choosing whether or not to enter the machine and uses (...) our choice (or rather the choice he assumes most people will have) as evidence against experientialist theories. But for this strategy to succeed it must be possible to distinguish between self-interested and non-self-interested reasons for declining to enter the machine, and there is no obvious way to do this without begging the question against the hedonist. In successive sections I then (a) consider a common misconception of Nozick’s conclusion (that he thinks machine life is the worst life), (b) consider different intuitions about what is important to well-being but missing from machine life, and finally (c) explain what “the experience requirement” is, and describe its relationship to debates about experientialist theories. (shrink)

Many of the computing systems programmed using Machine Learning are opaque: it is difficult to know why they do what they do or how they work. Explainable Artificial Intelligence aims to develop analytic techniques that render opaque computing systems transparent, but lacks a normative framework with which to evaluate these techniques’ explanatory successes. The aim of the present discussion is to develop such a framework, paying particular attention to different stakeholders’ distinct explanatory requirements. Building on an analysis of “opacity” (...) from philosophy of science, this framework is modeled after accounts of explanation in cognitive science. The framework distinguishes between the explanation-seeking questions that are likely to be asked by different stakeholders, and specifies the general ways in which these questions should be answered so as to allow these stakeholders to perform their roles in the Machine Learning ecosystem. By applying the normative framework to recently developed techniques such as input heatmapping, feature-detector visualization, and diagnostic classification, it is possible to determine whether and to what extent techniques from Explainable Artificial Intelligence can be used to render opaque computing systems transparent and, thus, whether they can be used to solve the Black Box Problem. (shrink)

A properly programmed artificially intelligent agent may eventually have one duty, the duty to satisfice expected welfare. We explain this claim and defend it against objections.

We propose a formal framework for interpretable machine learning. Combining elements from statistical learning, causal interventionism, and decision theory, we design an idealised explanation game in which players collaborate to find the best explanation for a given algorithmic prediction. Through an iterative procedure of questions and answers, the players establish a three-dimensional Pareto frontier that describes the optimal trade-offs between explanatory accuracy, simplicity, and relevance. Multiple rounds are played at different levels of abstraction, allowing the players to explore overlapping (...) causal patterns of variable granularity and scope. We characterise the conditions under which such a game is almost surely guaranteed to converge on a optimal explanation surface in polynomial time, and highlight obstacles that will tend to prevent the players from advancing beyond certain explanatory thresholds. The game serves a descriptive and a normative function, establishing a conceptual space in which to analyse and compare existing proposals, as well as design new and improved solutions. (shrink)

Artificial intelligence is a possibility that should not be ignored in any serious thinking about the future, and it raises many profound issues for ethics and public policy that philosophers ought to start thinking about. This article outlines the case for thinking that human-level machine intelligence might well appear within the next half century. It then explains four immediate consequences of such a development, and argues that machine intelligence would have a revolutionary impact on a wide range of (...) the social, political, economic, commercial, technological, scientific and environmental issues that humanity will face over the coming decades. (shrink)