This chapter proposed a novel design methodology called Value-Sensitive Design and its potential application to the field of artificial intelligence research and design. It discusses the imperatives in adopting a design philosophy that embeds values into the design of artificial agents at the early stages of AI development. Because of the high risk stakes in the unmitigated design of artificial agents, this chapter proposes that even though VSD may turn out to be a less-than-optimal design methodology, it currently (...) provides a framework that has the potential to embed stakeholder values and incorporate current design methods. The reader should begin to take away the importance of a proactive design approach to intelligent agents. (shrink)

While philosophers hold that it is patently absurd to blame robots or hold them morally responsible [1], a series of recent empirical studies suggest that people do ascribe blame to AI systems and robots in certain contexts [2]. This is disconcerting: Blame might be shifted from the owners, users or designers of AI systems to the systems themselves, leading to the diminished accountability of the responsible human agents [3]. In this paper, we explore one of the potential underlying reasons (...) for robot blame, namely the folk's willingness to ascribe inculpating mental states or "mens rea" to robots. In a vignette-based experiment (N=513), we presented participants with a situation in which an agent knowingly runs the risk of bringing about substantial harm. We manipulated agent type (human v. group agent v. AI-driven robot) and outcome (neutral v. bad), and measured both moral judgment (wrongness of the action and blameworthiness of the agent) and mental states attributed to the agent (recklessness and the desire to inflict harm). We found that (i) judgments of wrongness and blame were relatively similar across agent types, possibly because (ii) attributions of mental states were, as suspected, similar across agent types. This raised the question - also explored in the experiment - whether people attribute knowledge and desire to robots in a merely metaphorical way (e.g., the robot "knew" rather than really knew). However, (iii), according to our data people were unwilling to downgrade to mens rea in a merely metaphorical sense when given the chance. Finally, (iv), we report a surprising and novel finding, which we call the inverse outcome effect on robot blame: People were less willing to blame artificial agents for bad outcomes than for neutral outcomes. This suggests that they are implicitly aware of the dangers of overattributing blame to robots when harm comes to pass, such as inappropriately letting the responsible human agent off the moral hook. (shrink)

This paper focuses on the research field of machine ethics and how it relates to a technological singularity—a hypothesized, futuristic event where artificial machines will have greater-than-human-level intelligence. One problem related to the singularity centers on the issue of whether human values and norms would survive such an event. To somehow ensure this, a number of artificial intelligence researchers have opted to focus on the development of artificial moral agents, which refers to machines capable of moral reasoning, judgment, and (...) decision-making. To date, different frameworks on how to arrive at these agents have been put forward. However, there seems to be no hard consensus as to which framework would likely yield a positive result. With the body of work that they have contributed in the study of moral agency, philosophers may contribute to the growing literature on artificial moral agency. While doing so, they could also think about how the said concept could affect other important philosophical concepts. (shrink)

Legg and Hutter, as well as subsequent authors, considered intelligent agents through the lens of interaction with reward-giving environments, attempting to assign numeric intelligence measures to such agents, with the guiding principle that a more intelligent agent should gain higher rewards from environments in some aggregate sense. In this paper, we consider a related question: rather than measure numeric intelligence of one Legg- Hutter agent, how can we compare the relative intelligence of two Legg-Hutter agents? (...) We propose an elegant answer based on the following insight: we can view Legg-Hutter agents as candidates in an election, whose voters are environments, letting each environment vote (via its rewards) which agent (if either) is more intelligent. This leads to an abstract family of comparators simple enough that we can prove some structural theorems about them. It is an open question whether these structural theorems apply to more practical intelligence measures. (shrink)

Inspired by recent progress in multi-agent Reinforcement Learning (RL), in this work we examine the collective intelligent behaviour of theoretical universal agents by introducing a weighted mixture operation. Given a weighted set of agents, their weighted mixture is a new agent whose expected total reward in any environment is the corresponding weighted average of the original agents' expected total rewards in that environment. Thus, if RL agent intelligence is quantified in terms of performance across environments, the (...) weighted mixture's intelligence is the weighted average of the original agents' intelligences. This operation enables various interesting new theorems that shed light on the geometry of RL agent intelligence, namely: results about symmetries, convex agent-sets, and local extrema. We also show that any RL agent intelligence measure based on average performance across environments, subject to certain weak technical conditions, is identical (up to a constant factor) to performance within a single environment dependent on said intelligence measure. (shrink)

We define a notion of the intelligence level of an idealized mechanical knowing agent. This is motivated by efforts within artificial intelligence research to define real-number intelligence levels of compli- cated intelligent systems. Our agents are more idealized, which allows us to define a much simpler measure of intelligence level for them. In short, we define the intelligence level of a mechanical knowing agent to be the supremum of the computable ordinals that have codes the agent knows to (...) be codes of computable ordinals. We prove that if one agent knows certain things about another agent, then the former necessarily has a higher intelligence level than the latter. This allows our intelligence no- tion to serve as a stepping stone to obtain results which, by themselves, are not stated in terms of our intelligence notion (results of potential in- terest even to readers totally skeptical that our notion correctly captures intelligence). As an application, we argue that these results comprise evidence against the possibility of intelligence explosion (that is, the no- tion that sufficiently intelligent machines will eventually be capable of designing even more intelligent machines, which can then design even more intelligent machines, and so on). (shrink)

Interactions between an intelligent software agent and a human user are ubiquitous in everyday situations such as access to information, entertainment, and purchases. In such interactions, the ISA mediates the user’s access to the content, or controls some other aspect of the user experience, and is not designed to be neutral about outcomes of user choices. Like human users, ISAs are driven by goals, make autonomous decisions, and can learn from experience. Using ideas from bounded rationality, we frame these (...) interactions as instances of an ISA whose reward depends on actions performed by the user. Such agents benefit by steering the user’s behaviour towards outcomes that maximise the ISA’s utility, which may or may not be aligned with that of the user. Video games, news recommendation aggregation engines, and fitness trackers can all be instances of this general case. Our analysis facilitates distinguishing various subcases of interaction, as well as second-order effects that might include the possibility for adaptive interfaces to induce behavioural addiction, and/or change in user belief. We present these types of interaction within a conceptual framework, and review current examples of persuasive technologies and the issues that arise from their use. We argue that the nature of the feedback commonly used by learning agents to update their models and subsequent decisions could steer the behaviour of human users away from what benefits them, and in a direction that can undermine autonomy and cause further disparity between actions and goals as exemplified by addictive and compulsive behaviour. We discuss some of the ethical, social and legal implications of this technology and argue that it can sometimes exploit and reinforce weaknesses in human beings. (shrink)

Mandevillian intelligence is a specific form of collective intelligence in which individual cognitive shortcomings, limitations and biases play a positive functional role in yielding various forms of collective cognitive success. When this idea is transposed to the epistemological domain, mandevillian intelligence emerges as the idea that individual forms of intellectual vice may, on occasion, support the epistemic performance of some form of multi-agent ensemble, such as a socio-epistemic system, a collective doxastic agent, or an epistemic group agent. As a specific (...) form of collective intelligence, mandevillian intelligence is relevant to a number of debates in social epistemology, especially those that seek to understand how group (or collective) knowledge arises from the interactions between a collection of individual epistemic agents. Beyond this, however, mandevillian intelligence raises issues that are relevant to the research agendas of both virtue epistemology and applied epistemology. From a virtue epistemological perspective, mandevillian intelligence encourages us to adopt a relativistic conception of intellectual vice/virtue, enabling us to see how individual forms of intellectual vice may (sometimes) be relevant to collective forms of intellectual virtue. In addition, mandevillian intelligence is relevant to the nascent sub-discipline of applied epistemology. In particular, mandevillian intelligence forces us see the potential epistemic value of (e.g., technological) interventions that create, maintain or promote individual forms of intellectual vice. (shrink)

This paper investigates the nature of dispositional properties in the context of artificial intelligence systems. We start by examining the distinctive features of natural dispositions according to criteria introduced by McGeer (2018) for distinguishing between object-centered dispositions (i.e., properties like ‘fragility’) and agent-based abilities, including both ‘habits’ and ‘skills’ (a.k.a. ‘intelligent capacities’, Ryle 1949). We then explore to what extent the distinction applies to artificial dispositions in the context of two very different kinds of artificial systems, one based on (...) rule-based classical logic and the other on reinforcement learning. Here we defend three substantive claims. First, we argue that artificial systems are not equal in the kinds of dispositional properties they instantiate. In particular, we show that logical systems instantiate merely object-centered dispositions whereas reinforcement learning systems allow for the instantiation of agent-based abilities. Second, we explore the similarities and differences between the agent-centered abilities of artificial systems and those of humans, especially as relates to the important distinction made in the human case between habits and skills/intelligent capacities. The upshot is that the agent-centered abilities of truly intelligent artificial systems are distinctive enough to constitute a third type of agent-based ability — blended agent-based ability — raising substantial questions as to how we understand the nature of their agency. Third, we explore one aspect of this problem, focussing on whether systems of this type are properly considered ‘responsible agents’, at least in some contexts and for some purposes. The ramifications of our analysis will turn out to be directly relevant to various ethical concerns of artificial intelligence. (shrink)

Can an agent's intelligence level be negative? We extend the Legg-Hutter agent-environment framework to include punishments and argue for an affirmative answer to that question. We show that if the background encodings and Universal Turing Machine (UTM) admit certain Kolmogorov complexity symmetries, then the resulting Legg-Hutter intelligence measure is symmetric about the origin. In particular, this implies reward-ignoring agents have Legg-Hutter intelligence 0 according to such UTMs.

The objective of this essay is to provide the beginning of a principled classification of some of the ways space is intelligently used. Studies of planning have typically focused on the temporal ordering of action, leaving as unaddressed questions of where to lay down instruments, ingredients, work-in-progress, and the like. But, in having a body, we are spatially located creatures: we must always be facing some direction, have only certain objects in view, be within reach of certain others. How we (...) manage the spatial arrangement of items around us is not an afterthought: it is an integral part of the way we think, plan, and behave. The proposed classification has three main categories: spatial arrangements that simplify choice; spatial arrangements that simplify perception; and spatial dynamics that simplify internal computation. The data for such a classification is drawn from videos of cooking, assembly and packing, everyday observations in supermarkets, workshops and playrooms, and experimental studies of subjects playing Tetris, the computer game. This study, therefore, focuses on interactive processes in the medium and short term: on how agents set up their workplace for particular tasks, and how they continuously manage that workplace. (shrink)

In this paper I argue that many scholars involved in the contemporary free will debates have underappreciated the philosophical appeal of agent causation because the resources of contemporary emergentism have not been adequately introduced into the discussion. Whereas I agree that agent causation’s main problem has to do with its intelligibility, particularly with respect to the issue of how substances can be causally relevant, I argue that the notion of substance causation can be clearly articulated from an emergentist framework. According (...) to my proposal, a free agent is _a causally powerful substance that emerges in an anomic way from her constitutive mental events_, downwardly constraining, selecting and, in this way, having control on them. As we shall see, this particular concept of agent causation not only makes sense of the deep insight behind agent libertarianism, but it also provides us with the resources to solve some of the main objections that have been raised against it. It is true that here I cannot develop a complete defense of the evidential credentials of emergentism. Still, even if the considerations that follow do not serve to convince detractors of agent causal libertarian accounts of free will, they do suggest that libertarian agent causation is more promising than is typically acknowledged. (shrink)

Social machines are systems formed by technical and human elements interacting in a structured manner. The use of digital platforms as mediators allows large numbers of human participants to join such mechanisms, creating systems where interconnected digital and human components operate as a single machine capable of highly sophisticated behaviour. Under certain conditions, such systems can be described as autonomous and goal-driven agents. Many examples of modern Artificial Intelligence (AI) can be regarded as instances of this class of mechanisms. (...) We argue that this type of autonomous social machines has provided a new paradigm for the design of intelligent systems marking a new phase in the field of AI. The consequences of this observation range from methodological, philosophical to ethical. On the one side, it emphasises the role of Human-Computer Interaction in the design of intelligent systems, while on the other side it draws attention to both the risks for a human being and those for a society relying on mechanisms that are not necessarily controllable. The difficulty by companies in regulating the spread of misinformation, as well as those by authorities to protect task-workers managed by a software infrastructure, could be just some of the effects of this technological paradigm. (shrink)

This paper aims to justify the concept of natural intelligence in the biosemiotic context. I will argue that the process of life is (i) a cognitive/semiotic process and (ii) that organisms, from bacteria to animals, are cognitive or semiotic agents. To justify these arguments, the neural-type intelligence represented by the form of reasoning known as anthropic reasoning will be compared and contrasted with types of intelligence explicated by four disciplines of biology – relational biology, evolutionary epistemology, biosemiotics and the (...) systems view of life – not biased towards neural intelligence. The comparison will be achieved by asking questions related to the process of observation and the notion of true observers. To answer the questions I will rely on a range of established concepts including SETI (search for extraterrestrial intelligence), Fermi’s paradox, bacterial cognition, versions of the panspermia theory, as well as some newly introduced concepts including biocivilisations, cognitive/semiotic universes, and the cognitive/semiotic multiverse. The key point emerging from the answers is that the process of cognition/semiosis – the essence of natural intelligence – is a biological universal. (shrink)

Accountability is a cornerstone of the governance of artificial intelligence (AI). However, it is often defined too imprecisely because its multifaceted nature and the sociotechnical structure of AI systems imply a variety of values, practices, and measures to which accountability in AI can refer. We address this lack of clarity by defining accountability in terms of answerability, identifying three conditions of possibility (authority recognition, interrogation, and limitation of power), and an architecture of seven features (context, range, agent, forum, standards, process, (...) and implications). We analyse this architecture through four accountability goals (compliance, report, oversight, and enforcement). We argue that these goals are often complementary and that policy-makers emphasise or prioritise some over others depending on the proactive or reactive use of accountability and the missions of AI governance. (shrink)

The tendency to idealise artificial intelligence as independent from human manipulators, combined with the growing ontological entanglement of humans and digital machines, has created an “anthrobotic” horizon, in which data analytics, statistics and probabilities throw our agential power into question. How can we avoid the consequences of a reified definition of intelligence as universal operation becoming imposed upon our destinies? It is here argued that the fantasised autonomy of automated intelligence presents a contradistinctive opportunity for philosophical consciousness to understand itself (...) anew as holistic and co-creative, beyond the recent “analytic” moment of the history of philosophy. Here we introduce the concept of “crealectic intelligence”, a meta-analytic and meta-dialectic aspect of consciousness. Intelligent behaviour may consist in distinguishing discrete familiar parts or reproducible functions in the midst of noise via an analytic process of segmentation; intelligence may also manifest itself in the constitution of larger wholes and dynamic unities through a dialectic process of association or assemblage. But, by contrast, crealectic intelligence co-creates realities in the image of an ideal or truth, taking into account the desiring agent imbued with a sense of possibility, in a relationship not only with the Real but also with the creative sublime or “Creal”. (shrink)

The present narrative review seeks to unravel where we are now, and where we need to go to delineate the active ingredients in psychotherapy delivered by conversational agents (e.g., chatbots). While psychotherapy delivered by conversational agents has shown promising effectiveness for depression, anxiety, and psychological distress across several randomized controlled trials, little emphasis has been placed on the therapeutic processes in these interventions. The theoretical framework of this narrative review is grounded in prominent perspectives on the active ingredients (...) in psychotherapy. Based on search terms derived from various theoretical perspectives, we conducted a systematic literature search of four scientific databases and identified 17 studies. Across the selected studies, three themes emerged: relationship variables, emotional venting, and cognitive factors. While methodological issues compromise the epistemic value of this evidence base, core questions also remain to be answered. Such questions include, but are not limited to, whether humans can form affectionate bonds to inanimate beings and whether these kind of mental health treatments should be understood as psychotherapy or something else. Researchers should therefore be cautious when applying theories of psychotherapy in the realm of conversational agents. We conclude the paper by introducing recommendations for future research, which we hope will help instigate methodologically sound studies in this field. (shrink)

According to the Guise of the Good, an agent only does for a reason what she sees as good. One of the main motivations for the view is its apparent ability to explain why action for a reason must be intelligible to its agent, for on this view, an action is intelligible just in case it seems good. This motivation has come under criticism in recent years. Most notably, Kieran Setiya has argued that merely seeing one’s action as good does (...) not suffice to make the action intelligible. In this paper, I show that this objection has bite only because the Guise of the Good’s theory of intelligibility has yet seen little sustained articulation. Properly understood, this theory holds that an action is intelligible to an agent only if it appears to them to possess some substantive evaluative property. I then argue that this response to the objection has a significant implication for contemporary Guise of the Good theories, for it shows that the currently ascendant version of the theory, the attitudinal theory, cannot avail itself of the intelligibility motivation. (shrink)

In 2011, Hibbard suggested an intelligence measure for agents who compete in an adversarial sequence prediction game. We argue that Hibbard’s idea should actually be considered as two separate ideas: first, that the intelligence of such agents can be measured based on the growth rates of the runtimes of the competitors that they defeat; and second, one specific (somewhat arbitrary) method for measuring said growth rates. Whereas Hibbard’s intelligence measure is based on the latter growth-rate-measuring method, we survey (...) other methods for measuring function growth rates, and exhibit the resulting Hibbard-like intelligence measures and taxonomies. Of particular interest, we obtain intelligence taxonomies based on Big-O and Big-Theta notation systems, which taxonomies are novel in that they challenge conventional notions of what an intelligence measure should look like. We discuss how intelligence measurement of sequence predictors can indirectly serve as intelligence measurement for agents with Artificial General Intelligence (AGIs). (shrink)

Special Issue “Risks of artificial general intelligence”, Journal of Experimental and Theoretical Artificial Intelligence, 26/3 (2014), ed. Vincent C. Müller. http://www.tandfonline.com/toc/teta20/26/3# - Risks of general artificial intelligence, Vincent C. Müller, pages 297-301 - Autonomous technology and the greater human good - Steve Omohundro - pages 303-315 - - - The errors, insights and lessons of famous AI predictions – and what they mean for the future - Stuart Armstrong, Kaj Sotala & Seán S. Ó hÉigeartaigh - pages 317-342 - - (...) - The path to more general artificial intelligence - Ted Goertzel - pages 343-354 - - - Limitations and risks of machine ethics - Miles Brundage - pages 355-372 - - - Utility function security in artificially intelligent agents - Roman V. Yampolskiy - pages 373-389 - - - GOLEM: towards an AGI meta-architecture enabling both goal preservation and radical self-improvement - Ben Goertzel - pages 391-403 - - - Universal empathy and ethical bias for artificial general intelligence - Alexey Potapov & Sergey Rodionov - pages 405-416 - - - Bounding the impact of AGI - András Kornai - pages 417-438 - - - Ethics of brain emulations - Anders Sandberg - pages 439-457. (shrink)

This paper investigates the claim that artificial Intelligence Systems cannot be held morally responsible because they do not have an ability for agential self-awareness e.g. they cannot be aware that they are the agents of an action. The main suggestion is that if agential self-awareness and related first person representations presuppose an awareness of a self, the possibility of responsible artificial intelligence systems cannot be evaluated independently of research conducted on the nature of the self. Focusing on a specific (...) account of the self from the phenomenological tradition, this paper suggests that a minimal necessary condition that artificial intelligence systems must satisfy so that they have a capability for self-awareness, is having a minimal self defined as ‘a sense of ownership’. As this sense of ownership is usually associated with having a living body, one suggestion is that artificial intelligence systems must have similar living bodies so they can have a sense of self. Discussing cases of robotic animals as examples of the possibility of artificial intelligence systems having a sense of self, the paper concludes that the possibility of artificial intelligence systems having a ‘sense of ownership’ or a sense of self may be a necessary condition for having responsibility. (shrink)

The aim of this exploratory paper is to review an under-appreciated parallel between group agency and artificial intelligence. As both phenomena involve non-human goal-directed agents that can make a difference to the social world, they raise some similar moral and regulatory challenges, which require us to rethink some of our anthropocentric moral assumptions. Are humans always responsible for those entities’ actions, or could the entities bear responsibility themselves? Could the entities engage in normative reasoning? Could they even have rights (...) and a moral status? I will tentatively defend the (increasingly widely held) view that, under certain conditions, artificial intelligent systems, like corporate entities, might qualify as responsible moral agents and as holders of limited rights and legal personhood. I will further suggest that regulators should permit the use of autonomous artificial systems in high-stakes settings only if they are engineered to function as moral (not just intentional) agents and/or there is some liability-transfer arrangement in place. I will finally raise the possibility that if artificial systems ever became phenomenally conscious, there might be a case for extending a stronger moral status to them, but argue that, as of now, this remains very hypothetical. (shrink)

In their thought-provoking paper, Legg and Hutter consider a certain abstrac- tion of an intelligent agent, and define a universal intelligence measure, which assigns every such agent a numerical intelligence rating. We will briefly summarize Legg and Hutter’s paper, and then give a tongue-in-cheek argument that if one’s goal is to become more intelligent by cultivating music appreciation, then it is bet- ter to use classical music (such as Bach, Mozart, and Beethoven) than to use more recent pop (...) music. The same argument could be adapted to other media: books beat films, card games beat first-person shooters, parables beat dissertations, etc. We leave it to the reader to decide whether this argument tells us something about classical music, something about Legg-Hutter intelligence, or something about both. (shrink)

This chapter has two main goals: to update philosophers on the state of the art in the scientific psychology of intelligence, and to explain and evaluate challenges to the measurement invariance of intelligence tests. First, we provide a brief history of the scientific psychology of intelligence. Next, we discuss the metaphysics of intelligence in light of scientific studies in psychology and neuroimaging. Finally, we turn to recent skeptical developments related to measurement invariance. These have largely focused on attributability: Where do (...) the mechanisms and dispositions that explain people’s performance on tests of intelligence inhere – in the agent, in the local testing environment, in the culture, or in the interactions among these? After explaining what measurement invariance is in the context of intelligence testing, we explore the phenomenon of stereotype threat as a challenge to measurement invariance, as well as more recent work on overcoming or buffering against stereotype threat. (shrink)

Scientists and engineers seek to understand how real-world systems work and could work better. Any modeling method devised for such purposes must simplify reality. Ideally, however, the modeling method should be flexible as well as logically rigorous; it should permit model simplifications to be appropriately tailored for the specific purpose at hand. Flexibility and logical rigor have been the two key goals motivating the development of Agent-based Computational Economics (ACE), a completely agent-based modeling method characterized by seven specific modeling principles. (...) This perspective provides an overview of ACE, a brief history of its development, and its role within a broader spectrum of experiment-based modeling methods. (shrink)

Abstract: In this article I argue that the best way to understand the information turn is in terms of a fourth revolution in the long process of reassessing humanity's fundamental nature and role in the universe. We are not immobile, at the centre of the universe (Copernicus); we are not unnaturally distinct and different from the rest of the animal world (Darwin); and we are far from being entirely transparent to ourselves (Freud). We are now slowly accepting the idea that (...) we might be informational organisms among many agents (Turing), inforgs not so dramatically different from clever, engineered artefacts, but sharing with them a global environment that is ultimately made of information, the infosphere. (shrink)

Artificial agents (AAs), particularly but not only those in Cyberspace, extend the class of entities that can be involved in moral situations. For they can be conceived of as moral patients (as entities that can be acted upon for good or evil) and also as moral agents (as entities that can perform actions, again for good or evil). In this paper, we clarify the concept of agent and go on to separate the concerns of morality and responsibility of (...) agents (most interestingly for us, of AAs). We conclude that there is substantial and important scope, particularly in Computer Ethics, for the concept of moral agent not necessarily exhibiting free will, mental states or responsibility. This complements the more traditional approach, common at least since Montaigne and Descartes, which considers whether or not (artificial) agents have mental states, feelings, emotions and so on. By focussing directly on mind-less morality we are able to avoid that question and also many of the concerns of Artificial Intelligence. A vital component in our approach is the Method of Abstraction for analysing the level of abstraction (LoA) at which an agent is considered to act. The LoA is determined by the way in which one chooses to describe, analyse and discuss a system and its context. The Method of Abstraction is explained in terms of an interface or set of features or observables at a given LoA. Agenthood, and in particular moral agenthood, depends on a LoA. Our guidelines for agenthood are: interactivity (response to stimulus by change of state), autonomy (ability to change state without stimulus) and adaptability (ability to change the transition rules by which state is changed) at a given LoA. Morality may be thought of as a threshold defined on the observables in the interface determining the LoA under consideration. An agent is morally good if its actions all respect that threshold; and it is morally evil if some action violates it. That view is particularly informative when the agent constitutes a software or digital system, and the observables are numerical. Finally we review the consequences for Computer Ethics of our approach. In conclusion, this approach facilitates the discussion of the morality of agents not only in Cyberspace but also in the biosphere, where animals can be considered moral agents without their having to display free will, emotions or mental states, and in social contexts, where systems like organizations can play the role of moral agents. The primary cost of this facility is the extension of the class of agents and moral agents to embrace AAs. (shrink)

The first decade of this century has seen the nascency of the first mathematical theory of general artificial intelligence. This theory of Universal Artificial Intelligence (UAI) has made significant contributions to many theoretical, philosophical, and practical AI questions. In a series of papers culminating in book (Hutter, 2005), an exciting sound and complete mathematical model for a super intelligent agent (AIXI) has been developed and rigorously analyzed. While nowadays most AI researchers avoid discussing intelligence, the award-winning PhD thesis (Legg, (...) 2008) provided the philosophical embedding and investigated the UAI-based universal measure of rational intelligence, which is formal, objective and non-anthropocentric. Recently, effective approximations of AIXI have been derived and experimentally investigated in JAIR paper (Veness et al. 2011). This practical breakthrough has resulted in some impressive applications, finally muting earlier critique that UAI is only a theory. For the first time, without providing any domain knowledge, the same agent is able to self-adapt to a diverse range of interactive environments. For instance, AIXI is able to learn from scratch to play TicTacToe, Pacman, Kuhn Poker, and other games by trial and error, without even providing the rules of the games. These achievements give new hope that the grand goal of Artificial General Intelligence is not elusive. This article provides an informal overview of UAI in context. It attempts to gently introduce a very theoretical, formal, and mathematical subject, and discusses philosophical and technical ingredients, traits of intelligence, some social questions, and the past and future of UAI. (shrink)

This paper will address the moral implications of non-coercive interrogations in intelligence contexts. U.S. Army and CIA interrogation manuals define non-coercive interrogation as interrogation which avoids the use of physical pressure, relying instead on oral gambits. These methods, including some that involve deceit and emotional manipulation, would be mostly familiar to viewers of TV police dramas. As I see it, there are two questions that need be answered relevant to this subject. First, under what circumstances, if any, may a state (...) agent use deception or manipulation in the course of his or her duties? Second, if there are classes of persons who, by their activities, lose a legitimate expectation for honest-dealing, how are state agents to proceed when the identity of such persons is unclear? (shrink)

This paper proposes a metaphysical framework for distinguishing between human and machine intelligence. It posits two identical deterministic worlds -- one comprising a human agent and the other a machine agent. These agents exhibit different information processing mechanisms despite their apparent sameness in a causal sense. Providing a conceptual modeling of their difference, this paper resolves what it calls “the vantage point problem” – namely, how to justify an omniscient perspective through which a determinist asserts determinism from within the (...) supposedly deterministic universe. (shrink)

Social machines are systems formed by material and human elements interacting in a structured way. The use of digital platforms as mediators allows large numbers of humans to participate in such machines, which have interconnected AI and human components operating as a single system capable of highly sophisticated behavior. Under certain conditions, such systems can be understood as autonomous goal-driven agents. Many popular online platforms can be regarded as instances of this class of agent. We argue that autonomous social (...) machines provide a new paradigm for the design of intelligent systems, marking a new phase in AI. After describing the characteristics of goal-driven social machines, we discuss the consequences of their adoption, for the practice of artificial intelligence as well as for its regulation. (shrink)

Set aside fanciful doomsday speculations about AI. Even lower-level AIs, while otherwise friendly and providing us a universal basic income, would be able to do all our jobs. Also, we would over-rely upon AI assistants even in our personal lives. Thus, John Danaher argues that a human crisis of moral passivity would result However, I argue firstly that if AIs are posited to lack the potential to become unfriendly, they may not be intelligent enough to replace us in all (...) our jobs. If instead they are intelligent enough to replace us, the risk they become unfriendly increases, given that they would not need us and humans would just compete for valuable resources. Their hostility will not promote our moral passivity. Secondly, the use of AI assistants in our personal lives will become a problem only if we rely on them for almost all our decision-making and motivation. But such a (maximally) pervasive level of dependence raises the question of whether humans would accept it, and consequently whether the crisis of passivity will arise. (shrink)

As robots and Artificial Intelligences become more enmeshed in rich social contexts, it seems inevitable that we will have to make them into moral machines equipped with moral skills. Apart from the technical difficulties of how we could achieve this goal, we can also ask the ethical question of whether we should seek to create such Artificial Moral Agents (AMAs). Recently, several papers have argued that we have strong reasons not to develop AMAs. In response, we develop a comprehensive (...) analysis of the relevant arguments for and against creating AMAs, and we argue that all things considered we have strong reasons to continue to responsibly develop AMAs. The key contributions of this paper are threefold. First, to provide the first comprehensive response to the important arguments made against AMAs by Wynsberghe and Robbins (in “Critiquing the Reasons for Making Artificial Moral Agents”, Science and Engineering Ethics 25, 2019) and to introduce several novel lines of argument in the process. Second, to collate and thematise for the first time the key arguments for and against AMAs in a single paper. Third, to recast the debate away from blanket arguments for or against AMAs in general, to a more nuanced discussion about the use of what sort of AMAs, in what sort of contexts, and for what sort of purposes is morally appropriate. (shrink)

The following is an interpretation of Aquinas’ agent intellect focusing on Summa Theologiae I, qq. 75-89, and proposing that the agent intellect is a metaphysical rather than a formal a priori of human understanding. A formal a priori is responsible for the intelligibility as content of the object of human understanding and is related to Kant’s epistemological views; whereas a metaphysical a priori is responsible for intelligibility as mode of being of this same object. We can find in Aquinas’ text (...) many indications that the agent intellect is not productive of the universal as content but is, rather, productive of the abstracted or intelligible mode of being of the universal nature. This is because for Aquinas the universal nature, which is the object of human understanding, is present in the things themselves but with a different mode of being. Chapter 1 is intended to establish the fact which requires for Aquinas an agent intellect, and provides two very important principles: one is that the object of human understanding (the universal nature) is present in the things themselves and, the other, that it is not in the things themselves with a mode of being which makes it available to the intellectual eye. These two principles lead us to the main point of Chapter 2, namely the distinction between the intelligible object and its intelligible mode of being. Now, because knowing is receptive of the intelligible object (Chapter 3), which is present in the things themselves (Chapter 1), the agent intellect is productive not of the object’s intelligible content, but of its abstracted or intelligible mode of being (Chapter 4). (shrink)

In two experiments (total N=693) we explored whether people are willing to consider paintings made by AI-driven robots as art, and robots as artists. Across the two experiments, we manipulated three factors: (i) agent type (AI-driven robot v. human agent), (ii) behavior type (intentional creation of a painting v. accidental creation), and (iii) object type (abstract v. representational painting). We found that people judge robot paintings and human painting as art to roughly the same extent. However, people are much less (...) willing to consider robots as artists than humans, which is partially explained by the fact that they are less disposed to attribute artistic intentions to robots. (shrink)

One of the goals of physiologists who study the detailed physical, chemical,and neurological mechanisms operating within the human body is to understand the intricate causal processes which underlie human abilities and activities. It is doubtless premature to predict that they will eventually be able to explain the behaviour of a particular human being as we might now explain the behaviour of a pendulum clock or even the invisible changes occurring within the hardware of a modern electronic computer. Nonetheless, it seems (...) fair to say that hovering in the background of investigations into human physiology is the promise or threat, depending upon how one looks at the matter that human beings are complete physical-chemical systems and that all events taking place within their bodies and all movements of their bodies could be accounted for by physical causes if we but knew enough. I am not concerned at the moment with whether or not this ’mechanistic’ hypothesis is true, assuming that it is clear enough to be intelligible, nor with whether or not we could ever know that it is true. I wish to consider the somewhat more accessible yet equally important question whether our coming to believe that the hypothesis is true would warrant our relinquishing our conception of ourselves as beings who are capable of acting for reasons to achieve ends of our own choosing. I use the word ’warrant’ to indicate that I will not be discussing the possibility that believing the mechanistic hypothesis might lead us, as a matter of psychological fact, to think of human beings as mere automata, as objects whose movements are to be explained only by causes rather than by reasons, as are the actions of a personal subject. I intend to consider only whether the acceptance of mechanism would in fact justify such a change in conception. (shrink)

[Müller, Vincent C. (ed.), (2016), Fundamental issues of artificial intelligence (Synthese Library, 377; Berlin: Springer). 570 pp.] -- This volume offers a look at the fundamental issues of present and future AI, especially from cognitive science, computer science, neuroscience and philosophy. This work examines the conditions for artificial intelligence, how these relate to the conditions for intelligence in humans and other natural agents, as well as ethical and societal problems that artificial intelligence raises or will raise. The key issues (...) this volume investigates include the relation of AI and cognitive science, ethics of AI and robotics, brain emulation and simulation, hybrid systems and cyborgs, intelligence and intelligence testing, interactive systems, multi-agent systems, and superintelligence. Based on the 2nd conference on “Theory and Philosophy of Artificial Intelligence” held in Oxford, the volume includes prominent researchers within the field from around the world. (shrink)

This book is a collection of all the papers published in the special issue “Artificial Intelligence, Robots, and Philosophy,” Journal of Philosophy of Life, Vol.13, No.1, 2023, pp.1-146. The authors discuss a variety of topics such as science fiction and space ethics, the philosophy of artificial intelligence, the ethics of autonomous agents, and virtuous robots. Through their discussions, readers are able to think deeply about the essence of modern technology and the future of humanity. All papers were invited and (...) completed in spring 2020, though because of the Covid-19 pandemic and other problems, the publication was delayed until this year. I apologize to the authors and potential readers for the delay. I hope that readers will enjoy these arguments on digital technology and its relationship with philosophy. *** -/- Contents*** -/- Introduction : Descartes and Artificial Intelligence; Masahiro Morioka*** -/- Isaac Asimov and the Current State of Space Science Fiction : In the Light of Space Ethics; Shin-ichiro Inaba*** -/- Artificial Intelligence and Contemporary Philosophy : Heidegger, Jonas, and Slime Mold; Masahiro Morioka*** -/- Implications of Automating Science : The Possibility of Artificial Creativity and the Future of Science; Makoto Kureha*** -/- Why Autonomous Agents Should Not Be Built for War; István Zoltán Zárdai*** -/- Wheat and Pepper : Interactions Between Technology and Humans; Minao Kukita*** -/- Clockwork Courage : A Defense of Virtuous Robots; Shimpei Okamoto*** -/- Reconstructing Agency from Choice; Yuko Murakami*** -/- Gushing Prose : Will Machines Ever be Able to Translate as Badly as Humans?; Rossa Ó Muireartaigh***. (shrink)

Experts in Artificial Intelligence (AI) development predict that advances in the dvelopment of intelligent systems and agents will reshape vital areas in our society. Nevertheless, if such an advance isn't done with prudence, it can result in negative outcomes for humanity. For this reason, several researchers in the area are trying to develop a robust, beneficial, and safe concept of artificial intelligence. Currently, several of the open problems in the field of AI research arise from the difficulty of (...) avoiding unwanted behaviors of intelligent agents, and at the same time specifying what we want such systems to do. It is of utmost importance that artificial intelligent agents have their values aligned with human values, given the fact that we cannot expect an AI to develop our moral preferences simply because of its intelligence, as discussed in the Orthogonality Thesis. Perhaps this difficulty comes from the way we are addressing the problem of expressing objectives, values, and ends, using representational cognitive methods. A solution to this problem would be the dynamic cognitive approach proposed by Dreyfus, whose phenomenological philosophy defends that the human experience of being-in-the-world cannot be represented by the symbolic or connectionist cognitive methods. A possible approach to this problem would be to use theoretical models such as SED (situated embodied dynamics) to address the values learning problem in AI. (shrink)

Artificial intelligence (AI) is a digital technology that will be of major importance for the development of humanity in the near future. AI has raised fundamental questions about what we should do with such systems, what the systems themselves should do, what risks they involve and how we can control these. - After the background to the field (1), this article introduces the main debates (2), first on ethical issues that arise with AI systems as objects, i.e. tools made and (...) used by humans; here, the main sections are privacy (2.1), manipulation (2.2), opacity (2.3), bias (2.4), autonomy & responsibility (2.6) and the singularity (2.7). Then we look at AI systems as subjects, i.e. when ethics is for the AI systems themselves in machine ethics (2.8.) and artificial moral agency (2.9). Finally we look at future developments and the concept of AI (3). For each section within these themes, we provide a general explanation of the ethical issues, we outline existing positions and arguments, then we analyse how this plays out with current technologies and finally what policy conse-quences may be drawn. (shrink)

Papers from the conference on AI Risk (published in JETAI), supplemented by additional work. --- If the intelligence of artificial systems were to surpass that of humans, humanity would face significant risks. The time has come to consider these issues, and this consideration must include progress in artificial intelligence (AI) as much as insights from AI theory. -- Featuring contributions from leading experts and thinkers in artificial intelligence, Risks of Artificial Intelligence is the first volume of collected chapters dedicated to (...) examining the risks of AI. The book evaluates predictions of the future of AI, proposes ways to ensure that AI systems will be beneficial to humans, and then critically evaluates such proposals. 1 Vincent C. Müller, Editorial: Risks of Artificial Intelligence - 2 Steve Omohundro, Autonomous Technology and the Greater Human Good - 3 Stuart Armstrong, Kaj Sotala and Sean O’Heigeartaigh, The Errors, Insights and Lessons of Famous AI Predictions - and What they Mean for the Future - 4 Ted Goertzel, The Path to More General Artificial Intelligence - 5 Miles Brundage, Limitations and Risks of Machine Ethics - 6 Roman Yampolskiy, Utility Function Security in Artificially Intelligent Agents - 7 Ben Goertzel, GOLEM: Toward an AGI Meta-Architecture Enabling Both Goal Preservation and Radical Self-Improvement - 8 Alexey Potapov and Sergey Rodionov, Universal Empathy and Ethical Bias for Artificial General Intelligence - 9 András Kornai, Bounding the Impact of AGI - 10 Anders Sandberg, Ethics and Impact of Brain Emulations 11 Daniel Dewey, Long-Term Strategies for Ending Existential Risk from Fast Takeoff - 12 Mark Bishop, The Singularity, or How I Learned to Stop Worrying and Love AI -. (shrink)

We present a theory of cooperative-competitive intelligence (CCI), its measures, research program, and applications that stem from it. Within the framework of this theory, satisficing sub-optimal behavior is any behavior that does not promote a decrease in the prospective control of the functional action diversity/unpredictability (D/U) potential of the agent or team. This potential is defined as the entropy measure in multiple, context-dependent dimensions. We define the satisficing interval of behaviors as CCI. In order to manifest itself at individual or (...) team level, this capacity harnesses properties such as degeneracy, pleiotropy (pluri-potentiality), synergies, and metastability. Intelligence is embodied because intelligent behavior is deeply dependent on body functionalities, defined as entropy measures. We base our theory on three principles: (a) relativity of functional entropy/ information in agent (team)-environment systems, (b) tendency toward the satisficing level of D/U potential, and (c) tendency toward the non-decreasing D/U potential. The conjunction of these three principles provides existence of sub-optimal behaviors associated with CCI. First, we deal with the problem of how to reduce multidimensional behavior to a concept that accounts for the vast set of scenarios in which CCI is manifested. Secondly, we define and discuss the three interacting principles that underpin CCI behavior as well as providing an outline for a future CCI research program supported by agent￾based modeling and empirical research. Finally, we provide some preliminary practical issues that stem from the theory. (shrink)

Where artificial agents are not liable to be ascribed true moral agency and responsibility in their own right, we can understand them as acting as proxies for human agents, as making decisions on their behalf. What I call the ‘Moral Proxy Problem’ arises because it is often not clear for whom a specific artificial agent is acting as a moral proxy. In particular, we need to decide whether artificial agents should be acting as proxies for low-level (...) class='Hi'>agents — e.g. individual users of the artificial agents — or whether they should be moral proxies for high-level agents — e.g. designers, distributors or regulators, that is, those who can potentially control the choice behaviour of many artificial agents at once. Who we think an artificial agent is a moral proxy for determines from which agential perspective the choice problems artificial agents will be faced with should be framed: should we frame them like the individual choice scenarios previously faced by individual human agents? Or should we, rather, consider the expected aggregate effects of the many choices made by all the artificial agents of a particular type all at once? This paper looks at how artificial agents should be designed to make risky choices, and argues that the question of risky choice by artificial agents shows the moral proxy problem to be both practically relevant and difficult. (shrink)

Large-scale language technologies are increasingly used in various forms of communication with humans across different contexts. One particular use case for these technologies is conversational agents, which output natural language text in response to prompts and queries. This mode of engagement raises a number of social and ethical questions. For example, what does it mean to align conversational agents with human norms or values? Which norms or values should they be aligned with? And how can this be accomplished? (...) In this paper, we propose a number of steps that help answer these questions. We start by developing a philosophical analysis of the building blocks of linguistic communication between conversational agents and human interlocutors. We then use this analysis to identify and formulate ideal norms of conversation that can govern successful linguistic communication between humans and conversational agents. Furthermore, we explore how these norms can be used to align conversational agents with human values across a range of different discursive domains. We conclude by discussing the practical implications of our proposal for the design of conversational agents that are aligned with these norms and values. (shrink)

Virtue ethics has many times been suggested as a promising recipe for the construction of artificial moral agents due to its emphasis on moral character and learning. However, given the complex nature of the theory, hardly any work has de facto attempted to implement the core tenets of virtue ethics in moral machines. The main goal of this paper is to demonstrate how virtue ethics can be taken all the way from theory to machine implementation. To achieve this goal, (...) we critically explore the possibilities and challenges for virtue ethics from a computational perspective. Drawing on previous conceptual and technical work, we outline a version of artificial virtue based on moral functionalism, connectionist bottom–up learning, and eudaimonic reward. We then describe how core features of the outlined theory can be interpreted in terms of functionality, which in turn informs the design of components necessary for virtuous cognition. Finally, we present a comprehensive framework for the technical development of artificial virtuous agents and discuss how they can be implemented in moral environments. (shrink)

It has been argued that our currently most satisfactory social epistemology of science can’t account for science that is based on artificial intelligence (AI) because this social epistemology requires trust between scientists that can take full responsibility for the research tools they use, and scientists can’t take full responsibility for the AI tools they use since these systems are epistemically opaque. I think this argument overlooks that much AI-based science can be done without opaque models, and that agents can (...) take full responsibility for the systems they use even if these systems are opaque. Requiring that an agent fully understand how a system works is an untenably strong condition for that agent to take full responsibility for the system and risks absolving AI developers from responsibility for their products. AI-based science need not create trust-related social epistemological problems if we keep in mind that what makes both individual scientists and their use of AI systems trustworthy isn’t full transparency of the internal processing but their adherence to social and institutional norms that ensure that scientific claims can be trusted. (shrink)

Although virtue ethics has repeatedly been proposed as a suitable framework for the development of artificial moral agents, it has been proven difficult to approach from a computational perspective. In this work, we present the first technical implementation of artificial virtuous agents in moral simulations. First, we review previous conceptual and technical work in artificial virtue ethics and describe a functionalistic path to AVAs based on dispositional virtues, bottom-up learning, and top-down eudaimonic reward. We then provide the details (...) of a technical implementation in a moral simulation based on a tragedy of the commons scenario. The experimental results show how the AVAs learn to tackle cooperation problems while exhibiting core features of their theoretical counterpart, including moral character, dispositional virtues, learning from experience, and the pursuit of eudaimonia. Ultimately, we argue that virtue ethics provides a compelling path toward morally excellent machines and that our work provides an important starting point for such endeavors. (shrink)

-/- Les agents artificiels et les nouvelles technologies de l’information, de par leur capacité à établir de nouvelles dynamiques de transfert d’information, ont des effets perturbateurs sur les écosystèmes épistémiques. Se représenter la responsabilité pour ces chambardements représente un défi considérable : comment ce concept peut-il rendre compte de son objet dans des systèmes complexes dans lesquels il est difficile de rattacher l’action à un agent ou à une agente ? Cet article présente un aperçu du concept d’écosystème épistémique (...) et de la force de changement que représentent les nouvelles technologies pour celui-ci, puis illustre les difficultés rencontrées avec la notion classique de responsabilité dans les écosystèmes épistémiques. -/- Artificial agents and new information technologies, through their capacity to foster new information dynamics, have disruptive effects on epistemic ecosystems. Making sense of responsibility for these changes represent a challenge: the notion of responsibility struggles in complex systems, where actions are often hard to connect to an agent. This paper presents an overview of the concept of epistemic ecosystem and of the potential for disruption in new technologies, and then illustrates the difficulties with the classical notion of responsibility in such cases. (shrink)

A body schema is an agent's model of its own body that enables it to act on affordances in the environment. This paper presents a body schema system for the Learning Intelligent Decision Agent (LIDA) cognitive architecture. LIDA is a conceptual and computational implementation of Global Workspace Theory, also integrating other theories from neuroscience and psychology. This paper contends that the ‘body schema' should be split into three separate functions based on the functional role of consciousness in Global Workspace (...) Theory. There is (1) an online model of the agent's effectors and effector variables (Current Body Schema), (2) a long-term, recognitional storage of embodied capacities for action and affordances (Habitual Body Schema), and (3) "dorsal" stream information feeding directly from early perception to sensorimotor processes (Online Body Schema). This paper then discusses how the LIDA model of the body schema explains several experiments in psychology and ethology. (shrink)

This paper takes up a neglected problem for metaethical noncognitivism: the characterization of the acceptance states for agent-centered normative theories like Rational Egoism. If Egoism is a coherent view, the non-cognitivist needs a coherent acceptance state for it. This can be provided, as Dreier and Gibbard have shown. But those accounts fail when generalized, assigning the same acceptance state to normative theories that are clearly distinct, or assigning no acceptance state to theories that look to be intelligible. The paper makes (...) the case for this and then asks: What should we conclude if the problem cannot be solved? We might conclude that since Egoism is clearly a coherent view, the argument amounts to a refutation of noncognitivism. But we suggest another possibility. There is, on reflection, something incoherent, or at least odd, in standard formulations of Egoism; noncognitivism predicts this and so provides an intriguing explanation for this fact. (shrink)