This article proposes a conceptual mapping to outline salient properties and relations that allow for a knowledge transfer from the well-established greenwashing phenomenon to the more recent machinewashing. We account for relevant dissimilarities, indicating where conceptual boundaries may be drawn. Guided by a “reasoning by analogy” approach, the article addresses the structural analogy and machinewashing idiosyncrasies leading to a novel and theoretically informed model of machinewashing. Consequently, machinewashing is defined as a strategy that organizations adopt to engage in misleading behavior (...) about ethical Artificial Intelligence /algorithmic systems. Machinewashing involves misleading information about ethical AI communicated or omitted via words, visuals, or the underlying algorithm of AI itself. Furthermore, and going beyond greenwashing, machinewashing may be used for symbolic actions such as lobbying and prevention of stricter regulation. By outlining diverse theoretical foundations of the established greenwashing domain and their relation to specific research questions, the article proposes a machinewashing model and a set of theory-related research questions on the macro, meso, and micro-level for future machinewashing research. We conclude by stressing limitations and by outlining practical implications for organizations and policymakers. (shrink)

With automation of routine decisions coupled with more intricate and complex information architecture operating this automation, concerns are increasing about the trustworthiness of these systems. These concerns are exacerbated by a class of artificial intelligence that uses deep learning, an algorithmic system of deep neural networks, which on the whole remain opaque or hidden from human comprehension. This situation is commonly referred to as the black box problem in AI. Without understanding how AI reaches its conclusions, it is an open (...) question to what extent we can trust these systems. The question of trust becomes more urgent as we delegate more and more decision-making to and increasingly rely on AI to safeguard significant human goods, such as security, healthcare, and safety. Models that “open the black box” by making the non-linear and complex decision process understandable by human observers are promising solutions to the black box problem in AI but are limited, at least in their current state, in their ability to make these processes less opaque to most observers. A philosophical analysis of trust will show why transparency is a necessary condition for trust and eventually for judging AI to be trustworthy. A more fruitful route for establishing trust in AI is to acknowledge that AI is situated within a socio-technical system that mediates trust, and by increasing the trustworthiness of these systems, we thereby increase trust in AI. (shrink)

This article considers the issue of opacity as a problem for socially consequential mechanisms of classification and ranking, such as spam filters, credit card fraud detection, search engines, news trends, market segmentation and advertising, insurance or loan qualification, and credit scoring. These mechanisms of classification all frequently rely on computational algorithms, and in many cases on machine learning algorithms to do this work. In this article, I draw a distinction between three forms of opacity: opacity as intentional corporate or state (...) secrecy, opacity as technical illiteracy, and an opacity that arises from the characteristics of machine learning algorithms and the scale required to apply them usefully. The analysis in this article gets inside the algorithms themselves. I cite existing literatures in computer science, known industry practices, and do some testing and manipulation of code as a form of lightweight code audit. I argue that recognizing the distinct forms of opacity that may be coming into play in a given application is a key to determining which of a variety of technical and non-technical solutions could help to prevent harm. (shrink)

We are sceptical of concerns over the opacity of algorithmic decision tools. While transparency and explainability are certainly important desiderata in algorithmic governance, we worry that automated decision-making is being held to an unrealistically high standard, possibly owing to an unrealistically high estimate of the degree of transparency attainable from human decision-makers. In this paper, we review evidence demonstrating that much human decision-making is fraught with transparency problems, show in what respects AI fares little worse or better and argue that (...) at least some regulatory proposals for explainable AI could end up setting the bar higher than is necessary or indeed helpful. The demands of practical reason require the justification of action to be pitched at the level of practical reason. Decision tools that support or supplant practical reasoning should not be expected to aim higher than this. We cast this desideratum in terms of Daniel Dennett’s theory of the “intentional stance” and argue that since the justification of action for human purposes takes the form of intentional stance explanation, the justification of algorithmic decisions should take the same form. In practice, this means that the sorts of explanations for algorithmic decisions that are analogous to intentional stance explanations should be preferred over ones that aim at the architectural innards of a decision tool. (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)

The usefulness of machine learning algorithms has led to their widespread adoption prior to the development of a conceptual framework for making sense of them. One common response to this situation is to say that machine learning suffers from a “black box problem.” That is, machine learning algorithms are “opaque” to human users, failing to be “interpretable” or “explicable” in terms that would render categorization procedures “understandable.” The purpose of this paper is to challenge the widespread agreement about the existence (...) and importance of a black box problem. The first section argues that “interpretability” and cognates lack precise meanings when applied to algorithms. This makes the concepts difficult to use when trying to solve the problems that have motivated the call for interpretability. Furthermore, since there is no adequate account of the concepts themselves, it is not possible to assess whether particular technical features supply formal definitions of those concepts. The second section argues that there are ways of being a responsible user of these algorithms that do not require interpretability. In many cases in which a black box problem is cited, interpretability is a means to a further end such as justification or non-discrimination. Since addressing these problems need not involve something that looks like an “interpretation” of an algorithm, the focus on interpretability artificially constrains the solution space by characterizing one possible solution as the problem itself. Where possible, discussion should be reformulated in terms of the ends of interpretability. (shrink)

The algorithm, a building block of computer science, is defined from an intuitive and pragmatic point of view, through a methodological lens of philosophy rather than that of formal computation. The treatment extracts properties of abstraction, control, structure, finiteness, effective mechanism, and imperativity, and intentional aspects of goal and preconditions. The focus on the algorithm as a robust conceptual object obviates issues of correctness and minimality. Neither the articulation of an algorithm nor the dynamic process constitute the algorithm itself. Analysis (...) for implications in computer science and philosophy reveals unexpected results, new questions, and new perspectives on current questions, including the relationship between our informally construed algorithms and Turing machines. Exploration in terms of current computational and philosophical thinking invites further developments. (shrink)

Digital media and technologies have significantly transformed the ways we relate to the world, in the triple sense of Selbstwelt, Mitwelt, and Umwelt. Think of the quantification of the self, the number of followers and likes on social media, or using Google maps and similar tools to orient ourselves in a city, to find and choose a good restaurant, and so on. One might say that digital media and technologies have actually transformed our interpretation, understanding, and access to the world. (...) Now, if hermeneutics is the philosophy of interpretation, then we might suppose that hermeneutics should pay attention to these transformations. For us, digital hermeneutics is the study of the ways digital media and technologies mediate between humans and the world. It is also the study of the ways digital media and technologies are embedded in non-technological relations between humans and the world – psychological, social, cultural, and so on. (shrink)

This new translation is an extremely welcome addition to the continuing Cambridge Edition of Kant’s works. English-speaking readers of the third Critique have long been hampered by the lack of an adequate translation of this important and difficult work. James Creed Meredith’s much-reprinted translation has charm and elegance, but it is often too loose to be useful for scholarly purposes. Moreover it does not include the first version of Kant’s introduction, the so-called “First Introduction,” which is now recognized as indispensable (...) for an understanding of the work. Werner Pluhar’s more recent translation, which does include the First Introduction, is highly accurate when it confines itself to rendering Kant’s German. However, it is often more of a reconstruction than a translation, containing so many interpretative interpolations that it is often difficult to separate out Kant’s original text from the translator’s contributions. Paul Guyer and Eric Matthews have provided a translation that compares to or exceeds Pluhar’s in its literal approach to the German, but that confines all interpretative material to footnotes and endnotes, so that the text itself, with all its unclarities and ambiguities, lies open to view. In addition, Guyer, as editor of the volume, has provided a great deal of valuable supplementary material. This includes an introduction with an outline of the work and details of the history of its composition and publication, and a wealth of endnotes offering clarifications of the text, background information, and, most strikingly, many references to related passages in Kant’s voluminous writings, particularly in connection with Kant’s earlier writings related to aesthetics. The edition also records differences among the first three editions of the work, and—of particular interest—erasures from and additions to Kant’s manuscript of the First Introduction. Although the introduction and endnotes reflect interpretative views that are sometimes disputable, this supplementary material makes the present edition into a valuable resource even for those able to read the text in German. (shrink)

We only regard a system or a process as being “scientific” if it displays the three predominant characteristics of the natural sciences: predictability, repeatability and quantifiability. This by definition precludes intuition, subjective judgement, tacit knowledge, heuristics, dreams, etc. in other words, those attributes which are peculiarly human. Furthermore, this is resulting in a shift from judgment to calculation giving rise, in some cases, to an abject dependency on the machine and an inability to disagree with the outcome or even question (...) it. To tolerate such a situation could be seen as an abdication of professional responsibility. In complex technological and scientific environments, it is sometimes said that those who make best use of computers already know what the answer is (in ball park terms) before the calculation. (shrink)

In the spirit of Slavoj Žižek’s book, Hegel in a Wired Brain, this article asks how the questions central to Hans-Georg Gadamer’s philosophical hermeneutics are changed and complicated by the possibility of brain-to-brain communication and the datafication of thought made potential through brain-computer interfaces. By taking a phenomenological approach to understanding the nature of communication through a technology that does not require language for the transmission of ideas, this article explores how BCI communication confronts the ontological character of interpretation as (...) presented by Gadamer while also juxtaposing how the insights of Gadamer and Heidegger concerning temporality problematize materialist foundations of BCI communication. In the final analysis, it is argued that thought only becomes meaningful in accord with the temporality of Dasein and that hermeneutics remains central in BCI communication because the material replication of thought does not overcome the particular temporalities of historically effected consciousnesses. (shrink)

We argue that artificial networks are explainable and offer a novel theory of interpretability. Two sets of conceptual questions are prominent in theoretical engagements with artificial neural networks, especially in the context of medical artificial intelligence: Are networks explainable, and if so, what does it mean to explain the output of a network? And what does it mean for a network to be interpretable? We argue that accounts of “explanation” tailored specifically to neural networks have ineffectively reinvented the wheel. In (...) response to, we show how four familiar accounts of explanation apply to neural networks as they would to any scientific phenomenon. We diagnose the confusion about explaining neural networks within the machine learning literature as an equivocation on “explainability,” “understandability” and “interpretability.” To remedy this, we distinguish between these notions, and answer by offering a theory and typology of interpretation in machine learning. Interpretation is something one does to an explanation with the aim of producing another, more understandable, explanation. As with explanation, there are various concepts and methods involved in interpretation: Total or Partial, Global or Local, and Approximative or Isomorphic. Our account of “interpretability” is consistent with uses in the machine learning literature, in keeping with the philosophy of explanation and understanding, and pays special attention to medical artificial intelligence systems. (shrink)

Many of the mathematicians and scientists who guided the development of digital computers in the late 1940s, such as Alan Turing and John von Neumann, saw these new devices not just as tools for calculation but as devices that might employ the same principles as are exhibited in rational human thought. Thus, a subfield of what came to be called computer science assumed the label artificial intelligence. The idea of building artificial systems which could exhibit intelligent behavior comparable to that (...) of humans was a heady prospect, and the claims made on behalf of AI during the 1950s and 1960s were impossibly ambitious. Despite some theoretical and applied successes within the field, serious problems soon became evident. Instead of fulfilling the goal of quickly producing artificial intelligent agents which could compete with or outperform human beings, by the 1970s and 1980s AI had settled into a pattern of slower but real progress in modeling or simulating aspects of human intelligence. (shrink)

Unsupervised learning algorithms are widely used for many important statistical tasks with numerous applications in science and industry. Yet despite their prevalence, they have attracted remarkably little philosophical scrutiny to date. This stands in stark contrast to supervised and reinforcement learning algorithms, which have been widely studied and critically evaluated, often with an emphasis on ethical concerns. In this article, I analyze three canonical unsupervised learning problems: clustering, abstraction, and generative modeling. I argue that these methods raise unique epistemological and (...) ontological questions, providing data-driven tools for discovering natural kinds and distinguishing essence from contingency. This analysis goes some way toward filling the lacuna in contemporary philosophical discourse on unsupervised learning, as well as bringing conceptual unity to a heterogeneous field more often described by what it isnot(i.e., supervised or reinforcement learning) than by what itis. I submit that unsupervised learning is not just a legitimate subject of philosophical inquiry but perhaps the most fundamental branch of all AI. However, an uncritical overreliance on unsupervised methods poses major epistemic and ethical risks. I conclude by advocating for a pragmatic, error-statistical approach that embraces the opportunities and mitigates the challenges posed by this powerful class of algorithms. (shrink)

Artificial Intelligence in the form of different machine learning models is applied to Big Data as a way to turn data into valuable knowledge. The rhetoric is that ensuing predictions work well—with a high degree of autonomy and automation. We argue that we need to analyze the process of applying machine learning in depth and highlight at what point human knowledge production takes place in seemingly autonomous work. This article reintroduces classification theory as an important framework for understanding such seemingly (...) invisible knowledge production in the machine learning development and design processes. We suggest a framework for studying such classification closely tied to different steps in the work process and exemplify the framework on two experiments with machine learning applied to Facebook data from one of our labs. By doing so we demonstrate ways in which classification and potential discrimination take place in even seemingly unsupervised and autonomous models. Moving away from concepts of non-supervision and autonomy enable us to understand the underlying classificatory dispositifs in the work process and that this form of analysis constitutes a first step towards governance of artificial intelligence. (shrink)

In the last decades, supervised machine learning has seen the widespread growth of highly complex, non-interpretable models, of which deep neural networks are the most typical representative. Due to their complexity, these models have showed an outstanding performance in a series of tasks, as in image recognition and machine translation. Recently, though, there has been an important discussion over whether those non-interpretable models are able to provide any sort of understanding whatsoever. For some scholars, only interpretable models can provide understanding. (...) More popular, however, is the idea that understanding can come from a careful analysis of the dataset or from the model’s theoretical basis. In this paper, I wish to examine the possible forms of obtaining understanding of such non-interpretable models. Two main strategies for providing understanding are analyzed. The first involves understanding without interpretability, either through external evidence for the model’s inner functioning or through analyzing the data. The second is based on the artificial production of interpretable structures, through three main forms: post hoc models, hybrid models, and quasi-interpretable structures. Finally, I consider some of the conceptual difficulties in the attempt to create explanations for these models, and their implications for understanding. (shrink)

Computation and formalization are not modalities of pure abstractive operations. The essay tries to revise the assumption of the constitutive nonsensuality of the formal. The argument is that formalization is a kind of linear spatialization, which has significant visual dimensions. Thus, a connection can be discovered between visualization by figurative graphism and formalization by symbolic calculations: Both use spatial relations not only to represent but also to operate on epistemic, nonspatial, nonvisual entities. Descartes was one of the pioneers of using (...) this kind of two-dimensional spatiality as a cognitive instrument. (shrink)