Philosophy

Previous research in Explainable Artificial Intelligence (XAI) suggests that a main aim of explainability approaches is to satisfy specific interests, goals, expectations, needs, and demands regarding artificial systems (we call these 'stakeholders' desiderata') in a variety of contexts. However, the literature on XAI is vast, spreads out across multiple largely disconnected disciplines, and it often remains unclear how explainability approaches are supposed to achieve the goal of satisfying stakeholders' desiderata. This paper discusses the main classes of stakeholders calling for explainability of artificial systems and reviews their desiderata.

We argue that, to be trustworthy, Computational Intelligence (CI) has to do what it is entrusted to do for permissible reasons and to be able to give rationalizing explanations of its behavior which are accurate and graspable. We support this claim by drawing parallels with trustworthy human persons, and we show what difference this makes in a hypothetical CI hiring system. Finally, we point out two challenges for trustworthy CI and sketch a mechanism which could be used to generate sufficiently accurate as well as graspable rationalizing explanations for CI behavior.

We find ourselves surrounded by a rapidly increasing number of autonomous and semi-autonomous systems. Two grand challenges arise from this development: Machine Ethics and Machine Explainability. Machine Ethics, on the one hand, is concerned with behavioral constraints for systems, so that morally acceptable, restricted behavior results; Machine Explainability, on the other hand, enables systems to explain their actions and argue for their decisions in a way that human users can understand and justifiably trust them. In this paper, we try to motivate and work towards a framework combining Machine Ethics and Machine Explainability. Starting from a toy example, we detect various desiderata of such a framework and argue why they should and how they could be incorporated in autonomous systems. Our main idea is to apply a framework of formal argumentation theory both, for decision-making under ethically motivated constraints and for the task of generating useful explanations based on these constraints given only limited knowledge of the world. The result of our deliberations can be described as a first version of an ethically motivated, principle-governed framework combining Machine Ethics and Machine Explainability.

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.

National and international guidelines for trustworthy artificial intelligence (AI) consider explainability to be a central facet of trustworthy systems. This paper outlines a multidisciplinary rationale for explainability auditing. Specifically, we propose that explainability auditing can ensure the quality of explainability of systems in applied contexts and can be the basis for certification as a means to communicate whether systems meet certain explainability standards and requirements. Moreover, we emphasize that explainability auditing needs to take a multidisciplinary perspective, and we provide an overview of four perspectives (technical, psychological, ethical, legal) and their respective benefits with respect to explainability auditing.

Nowadays, users interact with applications in constantly changing environments. The plethora of I/O modalities is beneficial for a wide range of application areas such as virtual reality, cloud-based software, or scientific visualization. These areas require interfaces based not only on the traditional mouse and keyboard but also on gestures, speech, or highly-specialized and environment-dependent equipment. We introduce a hypergraph-based interaction model and its implementation as a distributed system, called MorphableUI. Its primary focus is to deliver a user- and developer-friendly way to establish dynamic connections between applications and interaction devices. We present an easy-to-use API for developers and a mobile frontend for users to set up their preferred interfaces. During runtime, MorphableUI transports interaction data between devices and applications. As one of the novelties, the system supports I/O transfer functions by automatically splitting, merging, and casting...

Im Zuge der Digitalisierung wird die Lebenswelt von Grundschulkindern mehr und mehr durch Informatiksysteme geprägt. Informatische Bildung in der Grundschule kann zur Entmystifizierung solcher Systeme maßgeblich beitragen und so die Grundlage für tiefgreifendes Verständnis legen. Bisher ist allerdings nur wenig bekannt über das Vorwissen der SchülerInnen in Bezug auf Informatiksysteme und über die Wirksamkeit der eingesetzten Unterrichtsmaterialien im Hinblick auf den Erwerb von informatischen Kompetenzen. In der hier beschriebenen Studie mit insgesamt 137 Kindern wurde das Vorwissen von Viertklässlern mit einem Leitfadeninterview ermittelt. Danach wurde eine fünfstündige Unterrichtseinheit zu algorithmischen Grundbausteinen und dem Mikrocontroller Calliope mini durchgeführt. Die Unterrichtseinheit endete mit einer Aufgabenstellung, für die die Kinder das Eingabe-Verarbeitung-Ausgabe-Prinzip (EVA-Prinzip) nachvollziehen und mit dem Mikrocontroller umsetzen mussten. Die Bearbeitung d...

We find ourselves surrounded by a rapidly increasing number of autonomous and semi-autonomous systems. Two grand challenges arise from this development: Machine Ethics and Machine Explainability. Machine Ethics, on the one hand, is concerned with behavioral constraints for systems, set up in a formal unambiguous, algorithmizable, and implementable way, so that morally acceptable, restricted behavior results; Machine Explainability, on the other hand, enables systems to explain their actions and argue for their decisions, so that human users can understand and justifiably trust them. In this paper, we stress the need to link and cross-fertilize these two areas. We point out how Machine Ethics calls for Machine Explainability, and how Machine Explainability involves Machine Ethics. We develop both these facets based on a toy example from the context of medical care robots. In this context, we argue that moral behavior, even if it were verifiable and verified, is not enough to establis...

Fernsehserien sind längst ein unabdingbarer Bestandteil unseres Alltags geworden. Insbesondere die Internetdienste haben durch erschwingliche Subskriptionsmodelle die Rezeption serieller Formate drastisch erhöht. Die mittlerweile nahezu uneingeschränkte Verfügbarkeit von Fernseh-und Internetserien und die erhöhte zeitliche und räumliche Flexibilität bei deren Rezeption wirken als Katalysatoren für diese Entwicklung. Wohl aufgrund ihrer Aktualität und Ubiquität erfreut sich die Fernsehserie auch im gegenwärtigen akademischen Diskurs einer großen Beliebtheit sowohl als Gegenstand der praxisbezogenen Lehre als auch der Forschung. Die Fernsehserienforschung-zuweilen in vielleicht zu vereinfachender Manier auch Fernsehforschung genannt 1-befindet sich, aus wissenschaftstheoretischer Perspektive betrachtet, aktuell in einer spannenden Phase, in der ihr Stellenwert als eigenständiger Untersuchungsbereich ausgehandelt wird. Im vorliegenden Beitrag werden Anhaltspunkte für die Beantwortung der Frage benannt, mit welchem Fug und Recht von ‚der Fernsehserienforschung' (im Singular) als einer universitären Disziplin die Rede sein kann. Um eine Antwort auf diese Fragestellung zu finden, ist es von Nöten, ein Kriterienset zu benennen, anhand dessen, intersubjektiv nachvollziehbar, eine wissenschaftssoziologische Standortbestimmung der Serienforschung erfolgen kann. Eine 10 Die Vielfalt der Begrifflichkeiten zur Bezeichnung der Pluralität der in einer Fernsehserie zur Anwendung kommenden Kommunikationsmittel ist bereits Ausdruck des Nebeneinanderforschens, in diesem Fall der Semiotik, Literatur-, Kultur-und Sprachwissenschaften.

We argue that explainable artificial intelligence (XAI), specifically reason-giving XAI, often constitutes the most suitable way of ensuring that someone can properly be held responsible for decisions that are based on the outputs of artificial intelligent (AI) systems. We first show that, to close moral responsibility gaps (Matthias 2004), often a human in the loop is needed who is directly responsible for particular AI-supported decisions. Second, we appeal to the epistemic condition on moral responsibility to argue that, in order to be responsible for her decision, the human in the loop has to have an explanation available of the system’s recommendation. Reason explanations are especially well-suited to this end, and we examine whether—and how—it might be possible to make such explanations fit with AI systems. We support our claims by focusing on a case of disagreement between human in the loop and AI system.

Informatics is the innovation driver of our time. From social media and artificial intelligence to autonomous cyber-physical systems: informatics-driven, digital products and services permeate our society in significant ways. Computer scientists, whether researchers or software developers, are shaping tomorrow's society. As a consequence, ethical, societal, and practical reasons demand that students of computer science and related subjects should receive at least a basic ethical education to be able to do justice to their ever-growing responsibilities and duties. Ethics for Nerds is an award-winning lecture that is being taught annually at Saarland University since 2016. The course has been continually updated and progressively improved over the years. In this paper, we share our experiences with and best practices for teaching the basics of ethics to students of computer science and offer advice on how to design a successful ethics course as part of a computer science study pro...

Designing trustworthy systems and enabling external parties to accurately assess the trustworthiness of these systems are crucial objectives. Only if trustors assess system trustworthiness accurately, they can base their trust on adequate expectations about the system and reasonably rely on or reject its outputs. However, the process by which trustors assess a system’s actual trustworthiness to arrive at their perceived trustworthiness remains underexplored. In this paper, we conceptually distinguish between trust propensity, trustworthiness, trust, and trusting behavior. Drawing on psychological models of assessing other people’s characteristics, we present the two-level Trustworthiness Assessment Model (TrAM). At the micro level, we propose that trustors assess system trustworthiness based on cues associated with the system. The accuracy of this assessment depends on cue relevance and availability on the system’s side, and on cue detection and utilization on the human’s side. At t...

This essay is concerned with the claim that when an agent acts for a normative reason, that normative reason is identical with his or her motivating reason. I will call this thesis the 'Identity Thesis'. Many philosophers treat the Identity Thesis as a commonplace or a truism, and some claim that it plays the role of a constraint on every theory of reasons (Dancy 2000). Accordingly, the Identity Thesis has been used to rule out certain ontological views about reasons. I argue that the Identity Thesis can be read in different ways and that each of its interpretations is controversial. My conclusion is that there is no convincing argument for the Identity Thesis and that there is a viable alternative to the Identity Thesis, namely, to think of the relation between motivating and normative reasons as one of representation, not identity. This allows for ontological views about reasons that are incompatible with the Identity Thesis. Therefore, the Identity Thesis need not be accepted as a truism or as a constraint on every theory of reasons.

Agents sometimes act for normative reasons—for reasons that  objectively favor their actions. Jill, for instance, calls a doctor for the normative reason that Kate is injured. In this article I explore a dispositional approach to acting for a normative reason. I argue for the need of epistemic, motivational, and executional dispositional elements of a theory of acting for a normative reason. Dispositions play a mediating role between, on the one hand, the normative reason and its normative force, and the action on the other hand. Thereby, they help to deal with problem cases such as cases of deviant causal chains and improper instrumental motivation.

In this paper I advocate a worldly account of normative reasons according to which there is an ontological gap between these and the premises of practical thought, i.e., motivating considerations. While motivating considerations are individuated fine-grainedly, normative reasons should be classified as coarse-grained entities, e.g., as states of affairs, in order to explain certain necessary truths about them and to make sense of how we count and weigh them. As I briefly sketch, acting for normative reasons is possible nonetheless if the connection between normative reasons and motivating considerations is a competence-based correspondence.

I argue for the viewthat there are important similarities between knowledge and acting for a normative reason. I interpret acting for a normative reason in terms of Sosa’s notion of an apt performance. Actions that are done for a normative reason are normatively apt actions. They are in accordance with a normative reason because
of a competence to act in accordance with normative reasons. I argue that, if Sosa’s account of knowledge as apt belief is correct, this means that acting for a normative reason is in many respects similar to knowledge. In order to strengthen Sosa’s account of knowledge, I propose to supplement it with an appeal to sub-competences. This clarifies howthis account can deal with certain Gettier cases, and it helps to understand how exactly acting for a normative reason is similar to apt belief.

Our virtuous heroes need not be moral philosophers. I may, for instance, admire a mariner who quit his job to spend his days pulling refugees out of the Mediterranean even though he never thought through highly sophisticated ethical theories, as I do for a living. But how do the virtuous master ethical challenges without sophisticated reflection on the right ethical theory, given that such reflection potentially reveals normative reasons for action, and urges us to take on new commitments? This puzzle is the starting point of Star's insightful inquiry into the nature of reasons and virtue, as well as into knowledge's role in acting well. Star aims to bridge the abyss between the ordinary deliberation of the virtuous and the highly sophisticated ethical theories that are relentlessly fine-tuned by philosophers in the Ivory Tower. In a refreshingly broad-minded manner, he highlights how normative ethics, metaethics, moral psychology, and epistemology are intertwined. Thereby, he reveals that philosophy is in dire need of a unifying story about the role(s) that normative reasons play in all these domains. Even though its significance might have been reinforced by evaluating more rivals to Star's view, this thought provoking book deserves to be widely noticed by philosophers, and it is to be hoped that it will do philosophy a great service by contributing to a better integration of various fields of research.

In his book Being Realistic about Reasons Tim Scanlon defends normative cognitivism – the view that normative attitudes are beliefs – against the objection that it is unable to explain how normative attitudes motivate agents to act. He provides a rational explanation of why agents act in accordance with normative attitudes. I suggest that this explanation relies on a dispositional structure and I argue that the dispositional structure can be detached from the rational theory it is combined with in the rational explanation. Thereby, the cognitivist’s answer to the motivational objection can be formulated as a dispositional explanation that is less controversial than the rational explanation and steers clear of disputes between rationalist internalists and externalists. Finally, I point out that there seems to be no rational explanation for actions performed for optional reasons, and that only dispositional explanations seem to accommodate these actions. This is a further advantage of giving dispositional explanations of acting on a normative judgement.
My aim is to strengthen the case for normative cognitivism by making it acceptable even to the instrumentalist, and by accommodating motivation by optional reasons. We may account for the fact that people are motivated by normative attitudes – in ways which, according to Tim’s (and my) views about rationality, are rationally required or rationally permitted – by giving an explanation that is uncommitted to this view of rationality.

Because normative reasons are non-psychological, and motivate us when we act for them, some have concluded that motivating reasons have to be non-psychological, too. This thought has served as part of an argument against the Causal-Psychological Account of Action, because the account states that motivating reasons are psychological. Recently, many authors recognized that there are two different notions of motivating reasons which are commonly being confused in this dispute. While the Anti-Psychologist is right that motivating reasons in her sense need to be classified as non-psychological, the proponent of the Causal-Psychological Account refers to something else by the term 'motivating reason' and rightly claims that it is psychological. However, this insight has not gained the uptake it deserves, partly because it has not been presented by way of a systematic evaluation of the different readings of the argument against the Causal-Psychological Account. This paper's aim is to point out more carefully the argument against the Causal-Psychological Account and the dilemma faced by that argument. This will hopefully show that and why motivating reasons – in the sense in which they figure in the Causal-Psychological Account – can be psychological even if agents act for non-psychological normative reasons.