Daniela Vacek

Although the phrase “responsible AI” is widely used in the AI industry, its meaning remains unclear. One can make sense of it indirectly, insofar as various notions of responsibility unproblematically attach to those involved in the creation and operation of AI technologies. It is less clear, however, whether the phrase makes sense when understood directly, that is, as the ascription of some sort of responsibility to AI systems themselves. This paper argues in the affirmative, drawing on a philosophically undemanding notion of role responsibility, and highlights the main consequences of this proposal for AI ethics.

The present paper tackles what we might call the AI achievement challenge, which has been the subject of recent debate in AI ethics. The challenge concerns the question of whether there are any achievement gaps due to artificial intelligence and what we should do if there are: how to fill them, or what policies are needed to decrease their impact on us if they cannot be filled. This paper argues that none of the proposed views is entirely satisfactory, even though they all have certain merits. The paper will provide two mutually compatible answers to the AI achievement challenge: one in terms of collective achievement and the other in terms of vicarious achievement.

Nolan (2014) successfully predicted a hyperintensional revolution in metaphysics and argued that hyperintensionality is not restricted to representations. However, it seems that one of the most promising candidates for non-representational hyperintensionality has not yet been considered as such: art. We argue that art can provide a rather strong case for non-representational hyperintensionality. One reason for this is that the hyperintensionality of art cannot be captured via a representational kind of hyperintensionality.

In this paper we present a tableau system for deontic logics with the operator of explicit permission. By means of this system the decidability of the considered logics can be proved. We will sketch how these logics are semantically defined by means of relating semantics and how they provide a simple solution to the free choice permission problem. In short, these logics employ relating implication and a certain propositional constant. These two are in turn used to define deontic operators similarly as in Andersonian-Kangerian reduction, which uses different intensional implications and constants.

Who is responsible when an AI machine causes something to go wrong? Or is there a gap in the ascription of responsibility? Answers range from claiming there is a unique responsibility gap, several different responsibility gaps, or no gap at all. In a nutshell, the problem is as follows: on the one hand, it seems fitting to hold someone responsible for a wrong caused by an AI machine; on the other hand, there seems to be no fitting bearer of responsibility for this wrong. In this article, we focus on a particular (aspect of the) AI responsibility gap: it seems fitting that someone should bear the legal consequences in scenarios involving AI machines with design defects; however, there seems to be no such fitting bearer. We approach this problem from the legal perspective, and suggest vicarious liability of AI manufacturers as a solution to this problem. Our proposal comes in two variants: the first one has a narrower range of application, but can be easily integrated in current legal frameworks; the second one requires a revision of current legal frameworks, but has a wider range of application. The latter variant employs a broadened account of vicarious liability. We emphasise strengths of the two variants and finally highlight how vicarious liability offers important insights for addressing a moral AI responsibility gap.

In this paper, we apply relating semantics to the widely discussed problem of free choice between permitted actions or situations in normative systems. Leaving aside contexts in which the free choice principle is obviously unacceptable (weak permission understood as an absence of prohibition) or uncontroversially valid (strong permission within systems with a closed repertoire of actions or situations), we concentrate on free choice for explicit permissions. In order to construct a formal representation of explicit permissions, we introduce a special constant, $\texttt {permit}$, which is analogous to the constant $\texttt {violation}$ used in the Andersonian–Kangerian approach to deontic logic with respect to prohibition and obligation. Consequently, we define a permission operator on the basis of the $\texttt {permit}$ constant and a relation of legitimization. The general idea is that $P \varphi $ is true if and only if $\texttt {permit}$ is true, which means that a permission is actually issued and $\varphi $ is legitimized by $\texttt {permit}$. The intuitive notion of legitimization is formally represented by an operator of relating implication: a non-classical implication that is semantically defined by adding a constraint to the classical meaning of implication to the effect that arguments are related by a special relation. The properties of this relation are based on an informal, intuitive meaning of legitimization and determine the properties of the permission operator. We show that the resulting permission operator possesses the free choice property and avoids certain unwanted consequences that follow from alternative approaches.

The received view in the debate on the form–content unity of poetry is that the possibility of paraphrase does not sit well with the unity conception. I will suggest a shift from paraphrase to translation, since the latter is substantially closer to the heart of the matter. I will heretically divert from the ‘commonplace’ view, which claims that poetry cannot be translated. However, I will argue that the possibility of translation in this sense can be reconciled, appearances notwithstanding, with the unity of form and content. A further surprising conclusion will be that, while this possibility prima facie appears to be the best argument against the unity, the contrary is the case.

Vicarious responsibility is a notoriously puzzling notion in normative reasoning. In this article we will explore two fundamental issues, which we will call the “explication problem” and the “justification problem”. The former issue concerns how vicarious responsibility can plausibly be defined in terms of other normative concepts. The latter issue concerns how ascriptions of vicarious responsibility can be justified. We will address these two problems by combining ideas taken from legal theory and moral philosophy. Our analysis will emphasise the importance of the voluntary involvement of the normative parties considered liable in a relation with other normative parties who causally contributed to a prohibited state-of-affairs.

We provide a fine-grained analysis of notions of regret and responsibility (such as agent-regret and individual responsibility) in terms of a language of multimodal logic. This language undergoes a detailed semantic analysis via two sorts of models: (i) relating models, which are equipped with a relation of propositional pertinence, and (ii) synonymy models, which are equipped with a relation of propositional synonymy. We specify a class of strictly relating models and show that each synonymy model can be transformed into an equivalent strictly relating model. Moreover, we define an axiomatic system that captures the notion of validity in the class of all strictly relating models.

In the present article we exploit the logical notions of correctness and completeness to provide an analysis of some fundamental problems that can be encountered by a software developer when transforming norms for traffic circulation into programming instructions. Relying on this analysis, we then introduce a question and answer procedure that can be helpful, in case of an accident, to clarify which components of an existing framework should be revised and to what extent software developers can be held responsible.

The present article provides a taxonomic analysis of bimodal logics of normative ideality and normative awfulness, two notions whose meaning is here explained in terms of the moral values pursued by a given community. Furthermore, the article addresses the traditional problem of a reduction among deontic concepts: we explore the possibility of defining other relevant normative notions, such as obligation, explicit permission and Hohfeldian relations, in terms of ideality and awfulness. Some proposals in this respect, which have been formulated in the literature over the years, are here improved and discussed with reference to the various logics that we will introduce

This paper formulates a new argument against the received view in the philosophy of poetry. The received view consists of three tenets: the unity of poetic form and poetic content; the impossibility of paraphrasing and translating poetry; and the hyperintensionality of poetry. We will explore the same detour via direct quotation that has been used by proponents of the received view. We will argue that the hyperintensionality and unity of quotation do not guarantee its untranslatability, and thus that the inference of untranslatability from hyperintensionality and unity is not universally valid. As a consequence, the claim that poetry cannot be translated is left without much support, since it relies on the same kind of inference. However, this does not mean that poetry is translatable; it means only that if it is not translatable, this is due to something other than the combination of hyperintensionality and unity.

Role realism is a promising realist theory of fictional names. Different versions of this theory have been suggested by Gregory Currie, Peter Lamarque, Stein Haugom Olsen, and Nicholas Wolterstorff. The general idea behind the approach is that fictional characters are to be analysed in terms of roles, which in turn can be understood as sets of properties. I will discuss several advantages and disadvantages of this approach. I will then propose a novel hyperintensional version of role realism, according to which fictional names are analysed in terms of individual concepts that cannot be matched by a reference. I will argue that this account avoids the main disadvantages of standard role realism.

In normative reasoning one typically refers to intervals of time across which norms are intended to hold, as well as to alternative possibilities representing hypothetical developments of a given scenario. Thus, deontic modalities are naturally intertwined with temporal and metaphysical ones. Furthermore, contemporary debates in philosophy suggest that a proper understanding of fundamental ethical principles, such as the Ought-Implies-Can thesis, requires a simultaneous analysis of these three families of concepts. In the present article we propose a general formal framework which allows for fine-grained multimodal reasoning in the normative domain. We provide an axiomatization for a novel system of propositional logic encoding the way in which possibilities and norms arising from different sources change over intervals of time. The usefulness of our framework is illustrated by analysing an ancient and particularly challenging `cold case', the Paradox of the Court.

Kulicki and Trypuz (2016) introduced three systems of multivalued deontic action logic to handle normative conflicts. The first system suggests a pessimistic view on normative conflicts, according to which any conflicting option represents something forbidden; the second system suggests an optimistic view, according to which any conflicting option represents something obligatory; finally, the third system suggests a neutral view, according to which any conflicting option represents something that is neither obligatory nor forbidden. The aim of the present paper is to propose a fourth system in this family, which comes with a realistic view on normative conflicts: a normative conflict remains unsolved unless it is generated by two or more normative sources that can be compared. In accordance with this, we will provide a more refined formal framework for the family of systems at issue, which allows for explicit reference to sources of norms. Conflict resolution is thus a consequence of a codified hierarchy of normative sources.

The present article is devoted to a logical treatment of some fundamental concepts involved in responsibility attribution. We specify a theoretical framework based on a language of temporal deontic logic with agent-relative operators for deliberate causal contribution. The framework is endowed with a procedure to solve normative conflicts which arise from the assessment of different normative sources. We provide a characterization result for a basic system within this framework and illustrate how the concepts formalized can be put at work in the analysis of examples of legal reasoning.

Are there fictional characters? Realists suggest that there are such entities, but these are non-concrete, non-actual or non-existent. Antirealists avoid this assumption by suggesting that fictional discourse is not to be taken at face value. However, any of these camps faces some serious troubles. This paper proposes a hybrid account that combines features of realism with features of antirealism. In particular, the semantic distinction between de dicto and de re is employed, and the resulting view suggests de dicto (role) realism and de re antirealism.

It is quite plausible to say that you may read or write implies that you may read and you may write (though possibly not both at once). This so-called free choice principle is well-known in deontic logic. Sadly, despite being so intuitive and seemingly innocent, this principle causes a lot of worries. The paper briefly but critically examines leading accounts of free choice permission present in the literature. Subsequently, the paper suggests to accept the free choice principle, but only as a default (or defeasible) rule, issuing to it a ticket-of-leave, granting it some freedom, until it commits an undesired inference.

Our aim is to examine ironical intertextual references and metalanguage in the movies of Woody Allen. The first part of the paper is devoted to intertextuality. To begin with, we specify the notion of intertextuality. Consequently, we provide certain typology of intertextual references, which takes its inspiration from Gérard Genette and Denis Fortin. Finally, we use this typology in investigating the phenomenon of intertextuality in Allen’s movies. The remaining part of the paper focuses on metalanguage. Again, we specify the crucial notion, the notion of metalanguage. This part is inspired mostly by Patricia Waugh’s work on metafiction. We introduce the distinction between embedded stories and metafiction. Subsequently, we examine their occurrences in Allen’s movies.

The paper examines two possible analyses of fictional names within Pavel Tichý’s Transparent Intensional Logic. The first of them is the analysis actually proposed by Tichý in his (1988) book The Foundations of Frege’s Logic. He analysed fictional names in terms of free variables. I will introduce, explain, and assess this analysis. Subsequently, I will explain Tichý’s notion of individual role (office, thing-to-be). On the basis of this notion, I will outline and defend the second analysis of fictional names. This analysis is close to the approach known in the literature as role realism (the most prominent advocates of this position are Nicholas Wolterstorff, Gregory Currie, and Peter Lamarque).