This paper explores the treatment of intelligent agents as innovations. Past writings in the area of intelligent agents focus on the technical merits and internal workings of agent-based solutions. By adopting a perspective on agents from an innovations point of view, a new and novel description of agents is put forth in terms of their degrees of innovativeness, competitive implications, and perceived characteristics. To facilitate this description, a series of innovation-based theoretical models are utilized (...) as a lens of analysis, namely Kleinschmidt and Cooper’s (J Prod Innovation Manage 8:240–251, 1991) market and technological newness map, Abernathy and Clark’s (Res Policy 14:3–22, 1985) competitive implications framework, and Moore and Benbasat’s (Inf Syst Res 2:192–222, 1991) list of perceived innovating characteristics. Together, these models provide a theoretical foundation by which to describe intelligent agents, yielding new insights and perceptions on this relatively new form of software application. (shrink)

The question of liability in the case of using intelligent agents is far from simple, and cannot sufficiently be answered by deeming the human user as being automatically responsible for all actions and mistakes of his agent. Therefore, this paper is specifically concerned with the significant difficulties which might arise in this regard especially if the technology behind software agents evolves, or is commonly used on a larger scale. Furthermore, this paper contemplates whether or not it is (...) possible to share the responsibility with these agents and what are the main objections surrounding the assumption of considering such agents as responsible entities. This paper, however, is not intended to provide the final answer to all questions and challenges in this regard, but to identify the main components, and provide some perspectives on how to deal with such issue. (shrink)

In How We Get Along, J. David Velleman argues for two related theses: first, that ‘making sense’ of oneself to oneself and others is a constitutive aim of action; second, that this fact about action grounds normativity. Examining each thesis in turn, I argue against the first that an agent may deliberately act in ways which make sense in terms of neither her self-conception nor others' conceptions of her. Against the second thesis, I argue that some vices are such that (...) the agents concerned would make more sense to neither themselves nor others if they were to reform, and, furthermore, that an agent may make more sense to herself and others by becoming more, rather than less, vicious. I conclude that both theses should be rejected. (shrink)

The emergence of intelligent software agents that operate autonomously with little or no human intervention has generated many doctrinal questions at a conceptual level and has challenged the traditional rules of contract especially those relating to the intention as an essential requirement of any contract conclusion. In this paper, we will try to explore some of these challenges, and shed light on the conflict between the traditional contract theory and the transactional practice in the case of using (...) class='Hi'>intelligent software agents. We will try further to examine how intelligent software agents differ from other software applications, and consider then how such differences are legally relevant. This paper, however, is not intended to provide the final answer to all questions and challenges in this regard, but to identify the main components, and provide perspectives on how to deal with such issue. (shrink)

This paper describes the layers of context leveraged by language-endowed intelligent agents (LEIAs) during incremental natural language understanding (NLU). Context is defined as a combination of (...

This paper instigates a discourse on the unplanned effects of intelligent agents in the context of their use on the Internet. By utilizing a social informatics framework as a lens of analysis, the study identifies several unanticipated consequences of using intelligent agents for information- and commerce-based tasks on the Internet. The effects include those that transpire over time at the organizational level, such as e-commerce transformation, operational encumbrance and security overload, as well as those that emerge (...) on a cultural level, such as trust affliction, skills erosion, privacy attrition and social detachment. Furthermore, three types of impacts are identified: economic, policy, and social. The discussion contends that economic impacts occur on the organizational level, social effects transpire on a cultural level, and policy impacts take place on both levels. These effects of the use of intelligent agents have seldom been predicted and discussed by visionaries, researchers, and practitioners in the field. The knowledge of these unplanned outcomes can improve our understanding of the overall impacts that innovative agent technologies may potentially have on organizations and individuals. Subsequently, this may help us develop better agent applications, facilitate the formulation of appropriate contingencies, and provide impetus for future research. (shrink)

Emotions play a significant role in human cognitive processes such as attention, motivation, learning, memory, and decision making. Many researchers have worked in the field of incorporating emotions in a cognitive agent. However, each model has its own merits and demerits. Moreover, most studies on emotion focus on steady-state emotions than emotion switching. Thus, in this article, a domain-independent computational model of emotions for intelligent agent is proposed that have modules for emotion elicitation, emotion regulation, and emotion transition. The (...) model is built on some well-known psychological theories such as appraisal theories of emotions, emotion regulation theory, and multistore human memory model. The design of the model is using the concept of fuzzy logic to handle uncertain and subjective information. The main focus is on primary emotions as suggested by Ekman; however, simultaneous elicitation of multiple emotions is also supported by the model. (shrink)

The perspective of modelling agents rather than using them for a specificed purpose entails a difference in approach. In particular an emphasis on veracity as opposed to efficiency. An approach using evolving populations of mental models is described that goes some way to meet these concerns. It is then argued that social intelligence is not merely intelligence plus interaction but should allow for individual relationships to develop between agents. This means that, at least, agents must be able (...) to distinguish, identify, model and address other agents, either individually or in groups. In other words that purely homogeneous interaction is insufficient. Two example models are described that illustrate these concerns, the second in detail where agents act and communicate socially, where this is determined by the evolution of their mental models. Finally some problems that arise in the interpretation of such simulations is discussed. (shrink)

Mental model ascription – also called mindreading – is the process of inferring the mental states of others, which happens as a matter of course in social interactions. But although ubiquitous, mindreading is presumably a highly variable process: people mindread to different extents and with _different results._ We hypothesize that human mindreading ability relies on a large number of personal and contextual features: the inherent abilities of specific individuals, their current physical and mental states, their knowledge of the domain of (...) discourse, their familiarity with the interlocutor, the risks associated with an incorrect assessment of intent, and so on. This paper presents a theory of mindreading that models diverse artificial intelligent agents using an inventory of parameters and value sets that represent traits of humans and features of discourse contexts. Examples are drawn from Maryland Virtual Patient, a prototype system that will permit medical trainees to diagnose and treat cognitively modeled virtual patients with the optional assistance of a virtual tutor. Since real patients vary greatly with respect to physiological and cognitive features, so must a society of virtual patients. Modeling such variation is one of the goals of the overall OntoAgent program of research and development. (shrink)

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)

As we move towards a world that is using machine learning and nascent artificial intelligence to analyse and, in many ways, guide most aspects of our lives, new forms of heterogeneous collaborative teams that include human/intelligent machine agents will become not just possible, but an inevitable part of our shared world. The conscious participation of the arts in the conversation about, and development and implementation of, these new collaborative possibilities is crucial, as the arts serve as our best (...) lens through which we can explore the full spectrum of potential possibilities and futures. Through the complex relationships formed through heterogeneous human/intelligent agent collaboration, we are given the opportunity to re-examine some of our assumptions and suppositions of teamwork and forms of relating and thinking within a group environment. (shrink)

Barring some civilisation-ending natural or man-made catastrophe, future scientists will likely incorporate fully fledged artificially intelligent agents in their ranks. Their tasks will include the conjecturing, extending and testing of hypotheses. At present human scientists have a number of methods to help them carry out those tasks. These range from the well-articulated, formal and unexceptional rules to the semi-articulated rules-of-thumb and intuitive hunches. If we are to hand over at least some of the aforementioned tasks to artificially (...) class='Hi'>intelligent agents, we need to find ways to make explicit and ultimately formal, not to mention computable, the more obscure of the methods that scientists currently employ with some measure of success in their inquiries. The focus of this talk is a problem for which the available solutions are at best semi-articulated and far from perfect. It concerns the question of how to conjecture new hypotheses or extend existing ones such that they do not save phenomena in gerrymandered or ad hoc ways. This talk puts forward a fully articulated formal solution to this problem by specifying what it is about the internal constitution of the content of a hypothesis that makes it gerrymandered or ad hoc. In doing so, it helps prepare the ground for the delegation of a full gamut of investigative duties to the artificially intelligent scientists of the future. (shrink)

Artificial social intelligence (ASI) agents have great potential to aid the success of individuals, human–human teams, and human–artificial intelligence teams. To develop helpful ASI agents, we created an urban search and rescue task environment in Minecraft to evaluate ASI agents’ ability to infer participants’ knowledge training conditions and predict participants’ next victim type to be rescued. We evaluated ASI agents’ capabilities in three ways: (a) comparison to ground truth—the actual knowledge training condition and participant actions; (b) (...) comparison among different ASI agents; and (c) comparison to a human observer criterion, whose accuracy served as a reference point. The human observers and the ASI agents used video data and timestamped event messages from the testbed, respectively, to make inferences about the same participants and topic (knowledge training condition) and the same instances of participant actions (rescue of victims). Overall, ASI agents performed better than human observers in inferring knowledge training conditions and predicting actions. Refining the human criterion can guide the design and evaluation of ASI agents for complex task environments and team composition. (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)

El 28 de mayo de 1970, Jorge Luis Borges dictó la conferencia “Junín y la Conquista del Desierto” en esa la ciudad. Entre el público presente, un agente de los servicios de inteligencia local, escuchó, escribió y envió a sus superiores una versión de la conferencia y sus impresiones, registro singular que consta en el Archivo de la Dirección de Inteligencia de la Policía de la Provincia de Buenos Aires (DIPPBA). El objetivo del artículo es inscribir el documento en dos (...) niveles: (1) exponer algunas notas y reflexiones acerca de la narración de vidas, momentos biográficos y autobiográficos en la escritura y la estética literaria de Borges que señalan temas afines a lo supuestamente expuesto en la conferencia; (2) ubicar el documento, las lógicas del archivo y la coyuntura política que lo rodeaba en diálogo con lo anterior. Y, entretejiendo ambos, analizar en términos metodológicos y analíticos las relaciones tensionadas entre biografía, autobiografía y archivo. (shrink)

In this paper, we use some elements of the philosophical theories of Wilfrid Sellars and Robert Brandom for examining the interactions between humans and machines. In particular, we adopt the concept of the space of reasons for analyzing the status of artificial intelligent agents. One could argue that AIAs, like the widely used recommendation systems, have already entered the space of reasons, since they seem to make knowledge claims that we use as premises for further claims. This, in (...) turn, can lead to a sense of alienation because AIAs do not quite play by the rules of the space of the reason. We, therefore, ask somewhat pointedly whether aliens have entered the space of reasons. A closer look reveals that it is a misconception to consider AIAs as being in the space of reasons. In fact, they should be seen as very sophisticated tools. Since these tools affect our own acting in the space of reasons, special regulations are required for their proper use. (shrink)

In the past decade, the fields of machine learning and artificial intelligence (AI) have seen unprecedented developments that raise human-machine interactions (HMI) to the next level.Smart machines, i.e., machines endowed with artificially intelligent systems, have lost their character as mere instruments. This, at least, seems to be the case if one considers how humans experience their interactions with them. Smart machines are construed to serve complex functions involving increasing degrees of freedom, and they generate solutions not fully anticipated by (...) humans. Consequently, their performances show a touch of action and even autonomy. HMI is therefore often described as a sort of “cooperation” rather than as a mere application of a tool. Some authors even go as far as subsuming cooperation with smart machines under the label ofpartnership, akin to cooperation between human agents sharing a common goal. In this paper, we explore how far the notion of shared agency and partnership can take us in our understanding of human interaction with smart machines. Discussing different topoi related to partnerships in general, we suggest that different kinds of “partnership” depending on the form of interaction between agents need to be kept apart. Building upon these discussions, we propose a tentative taxonomy of different kinds of HMI distinguishing coordination, collaboration, cooperation, and social partnership. (shrink)

The present paper uses a Flourishing Ethics analysis to address the question of which ethical values and principles should be “instilled” into artificially intelligent agents. This is an urgent question that is still being asked seven decades after philosopher/scientist Norbert Wiener first asked it. An answer is developed by assuming that human flourishing is the central ethical value, which other ethical values, and related principles, can be used to defend and advance. The upshot is that Flourishing Ethics can (...) provide a common underlying ethical foundation for a wide diversity of cultures and communities around the globe; and the members of each specific culture or community can add their own specific cultural values—ones which they treasure, and which help them to make sense of their moral lives. (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)

Intelligent Transportation Systems are expected to automate how parking slots are booked by trucks. The intrinsic dynamic nature of this problem, the need of explanations and the inclusion of private data justify an agent-based solution. Agents solving this problem act with a Believe Desire Intentions reasoning, and are implemented with JASON. Privacy of trucks becomes protected sharing a list of parkings ordered by preference. Furthermore, the process of assigning parking slots takes into account legal requirements on breaks and (...) driving time limits. Finally, the agent simulations use the distances, the number of trucks and parkings corresponding to the proportions of the current European Union data. The performance of the proposed solution is tested in these simulations with three different distances against an alternative with complete knowledge. The difference in efficiency, the number of illegal breaks and the traveled distances are measured in them. Comparing the results, we can conclude that the nonprivate alternative is slightly better in performance while both alternatives do not produce illegal breaks. In this way the simulations show that the proposed privacy protection does not impose a relevant handicap in efficiency. (shrink)

Conversational artificial intelligence (CAI) presents many opportunities in the psychotherapeutic landscape—such as therapeutic support for people with mental health problems and without access to care. The adoption of CAI poses many risks that need in-depth ethical scrutiny. The objective of this paper is to complement current research on the ethics of AI for mental health by proposing a holistic, ethical, and epistemic analysis of CAI adoption. First, we focus on the question of whether CAI is rather a tool or an (...) agent. This question serves as a framework for the subsequent ethical analysis of CAI focusing on topics of (self-) knowledge, (self-)understanding, and relationships. Second, we propose further conceptual and ethical analysis regarding human-AI interaction and argue that CAI cannot be considered as an equal partner in a conversation as is the case with a human therapist. Instead, CAI’s role in a conversation should be restricted to specific functions. (shrink)

The necessity to model the mental ingredients of norm compliance is a controversial issue within the study of norms. So far, the simulation-based study of norm emergence has shown a prevailing tendency to model norm conformity as a thoughtless behavior, emerging from social learning and imitation rather than from specific, norm-related mental representations. In this paper, the opposite stance—namely, a view of norms as hybrid, two-faceted phenomena, including a behavioral/social and an internal/mental side—is taken. Such a view is aimed at (...) accounting for the difference between norms, on one hand, and either behavioral regularities (conventions) on the other. This paper, in particular, is addressed to find out the internal ingredients required for the former distinction, i.e., to model norms as distinct from mere conventions, and defined as behaviors spreading to the extent that and because the corresponding commands and beliefs do spread as well. After a brief presentation of a normative agent architecture, the results of agent-based simulations testing the impact of norm recognition and the role of normative beliefs in the emergence and innovation of social norms are presented and discussed. More specifically, the present work will endeavour to show that a sudden external constraint (e.g. a barrier preventing agents from moving among social settings) facilitates norm innovation: under such a condition, agents provided with a module for telling what a norm is can generate new (social) norms by forming new normative beliefs, irrespective of the most frequent actions. (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)

Artificial Intelligence is supposed to perform tasks autonomously, make competent decisions, and interact socially with people. From a psychological perspective, AI can thus be expected to impact users’ three Basic Psychological Needs, namely autonomy, competence, and relatedness to others. While research highlights the fulfillment of these needs as central to human motivation and well-being, their role in the acceptance of AI applications has hitherto received little consideration. Addressing this research gap, our study examined the influence of BPN Satisfaction on Intention (...) to Use an AI assistant for personal banking. In a 2×2 factorial online experiment, 282 participants watched a video of an AI finance coach with a female or male synthetic voice that exhibited either high or low agency. In combination, these factors resulted either in AI assistants conforming to traditional gender stereotypes or in non-conforming conditions. Although the experimental manipulations had no significant influence on participants’ relatedness and competence satisfaction, a strong effect on autonomy satisfaction was found. As further analyses revealed, this effect was attributable only to male participants, who felt their autonomy need significantly more satisfied by the low-agency female assistant, consistent with stereotypical images of women, than by the high-agency female assistant. A significant indirect effects model showed that the greater autonomy satisfaction that men, unlike women, experienced from the low-agency female assistant led to higher ITU. The findings are discussed in terms of their practical relevance and the risk of reproducing traditional gender stereotypes through technology design. (shrink)

This study evaluated how call center agents manage their emotions when interacting with customers with different emotional states. The coping mechanisms employees develop through experience can impact their communication and satisfaction with customer service. A study was conducted using a descriptive-correlational design in three Business Process Outsourcing companies in Cebu City, Philippines. The study aimed to determine employees' agreement and effectiveness in self-awareness, self-management, social awareness, and relationship management. An online sample size calculator was used to gather data, and (...) 150 respondents were chosen out of 243 call center agents. The data was analyzed using simple frequency, percentage, weighted mean, and ANOVA type III. The study indicates that age is crucial in the respondents' self-management (F ratio= 3,146, p= .412). This suggests that BPO employees become better at managing themselves as they get older. Additionally, the study respondents' monthly income significantly impacts their social awareness level (F ratio= 5,144, p= .040). This means that those with higher incomes tend to have greater social awareness than those with lower incomes. The BPO industry should prioritize candidates with high emotional intelligence during hiring. They can better adapt to the fast-paced work environment. Graveyard shift workers should receive higher wages and benefits. Employees should attend workshops on management, leadership, and emotional intelligence. Annual team-building events should also be considered for a positive work environment. (shrink)

In this paper we introduce a new generation of language trainers: intelligent virtual agents (IVAs) with human appearance and the capability to teach foreign language vocabulary. We report results from studies that we have conducted with Billie, an IVA employed as a vocabulary trainer, as well as research findings on the acceptance of the agent as a trainer by adults and children. The results show that Billie can train humans as well as a human teacher can and that (...) both adults and children accept the IVA as a trainer. The advantages of IVAs are multiple. First, their teaching methods can be based on neuropsychological research findings concerning memory and learning practice. Second, virtual teachers can provide individualized training. Third, they coach users during training, are always supportive, and motivate learners to train. Fourth, agents will reside in the user's mobile devices and thus be at the user's disposal everywhere and anytime. Agents in apps will make foreign language training accessible to anybody at low cost. This will enable people around the world, including physically, financially, and geographically disadvantaged persons, to learn a foreign language and help to facilitate multilingualism. (shrink)

The doctrine of agent-causation has been suggested by many interested in defending libertarian theories of free action to provide the conceptual apparatus necessary to make the notion of incompatibility freedom intelligible. In the present essay the conceptual viability of the doctrine of agent-causation will be assessed. It will be argued that agent-causation is, insofar as it is irreducible to event-causation, mysterious at best, totally unintelligible at worst. First, the arguments for agent-causation made by such eighteenth-century luminaries as Samuel Clarke and (...) Thomas Reid will be considered alongside the defenses of agent-causation proffered in this century by C.A. Campbell, Roderick Chisholm, and Richard Taylor. It will be shown that the case for agent-causation made by these figures is ultimately unconvincing. Two defenses of agent-causation made within the past ten years will then be taken up for examination and critique. First, Timothy O'Connor's attempt at advancing an unrefined and unrepentant doctrine of agent-causation will be shown to suffer from the same maladies as its predecessors. Next, Randolph Clarke's causal agent-causal theory of free action, which seeks a via media between agent-causal theories of free action and causal theories of action, is examined. Clarke's theory is an attempt at providing an account of how both events and agents qua substances can be the codeterminants of free actions. Despite the improvement of Clarke's theory over more conventional agent-causal theories of free action, it will be shown that agent-causation makes his theory more cumbersome than it needs to be. Clarke is able to get as much mileage out of a causal indeterminacy theory of action that does not require him to posit obscure agent-causes. Finally, a sketch of an alternative theory of free action will be offered. While it may suffer from its own conceptual deficiencies, it may not suffer from the same conceptual problems as agent-causal theories of free action. (shrink)

Purpose The purpose of this paper is to demonstrate the advancements in artificial intelligence (AI) and conversational agents, emphasizing their potential benefits while also highlighting the need for vigilant monitoring to prevent unethical applications. Design/methodology/approach As AI becomes more prevalent in academia and research, it is crucial to explore ways to ensure ethical usage of the technology and to identify potentially unethical usage. This manuscript uses a popular AI chatbot to write the introduction and parts of the body of (...) a manuscript discussing conversational agents, the ethical usage of chatbots and ethical concerns for academic researchers. Findings The authors reveal which sections were written entirely by the AI using a conversational agent. This serves as a cautionary tale highlighting the importance of ethical considerations for researchers and students when using AI and how educators must be prepared for the increasing prevalence of AI in the academy and industry. Measures to mitigate potential unethical use of this evolving technology are also discussed in the manuscript. Originality/value As conversational agents and chatbots increase in society, it is crucial to understand how they will impact the community and how we can live with technology instead of fighting against it. (shrink)

This paper urges that if we wish to give social intelligence to our agents, it pays to look at how we acquired our social intelligence ourselves. We are born with drives and motives that are innate and deeply social. Next, as children we are socialized to acquire norms and values and to understand rituals large and small. These social elements are the core of our being. We capture them in the acronym GRASP: Groups, Rituals, Affiliation, Status, Power. As a (...) consequence, economic rationality or logical reasoning do not suffice when it comes to social intelligence. Basic features of our social behaviour, of the kind that one sees early in the lives of children, need to be prominent. These include fear, love, and aggression. They extend to the combination of these drives with basic social clues from the environment such as big and small, good and bad, as well as culture-specific specializations of these. This will make agents respond differentially to inferred attributes such as gender, age, group membership. This level of universality in social intelligence should receive our full attention. The general insights gained can then be re-used in myriad implementations to specific modelling issues. (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)

Deep learning (DL) has become increasingly central to science, primarily due to its capacity to quickly, efficiently, and accurately predict and classify phenomena of scientific interest. This paper seeks to understand the principles that underwrite scientists’ epistemic entitlement to rely on DL in the first place and argues that these principles are philosophically novel. The question of this paper is not whether scientists can be justified in trusting in the reliability of DL. While today’s artificial intelligence exhibits characteristics common to (...) both scientific instruments and scientific experts, this paper argues that the familiar epistemic categories that justify belief in the reliability of instruments and experts are distinct, and that belief in the reliability of DL cannot be reduced to either. Understanding what can justify belief in AI reliability represents an occasion and opportunity for exciting, new philosophy of science. (shrink)

Much of our everyday, embodied action comes in the form of smooth coping. Smooth coping is skillful action that has become habituated and ingrained, generally placing less stress on cognitive load than considered and deliberative thought and action. When performed with skill and expertise, walking, driving, skiing, musical performances, and short-order cooking are all examples of the phenomenon. Smooth coping is characterized by its rapidity and relative lack of reflection, both being hallmarks of automatization. Deliberative and reflective actions provide the (...) contrast case. In Dreyfus’ classic view, smooth coping is “mindless” absorption into action, being in the flow, and any reflective thought will only interrupt this flow. Building on the pragmatist account of Dewey, others, such as Sutton, Montero, and Gallagher, insist on the intelligent flexibility built into smooth coping, suggesting that it is not equivalent to automatization. We seek to answer two complementary challenges in this article. First, how might we model smooth coping in autonomous agents at fine granularity? Second, we use this model of smooth coping to show how we might implement smooth coping in artificial intelligent agents. We develop a conceptual model of smooth coping in LIDA. LIDA is an embodied cognitive architecture implementing the global workspace theory of consciousness, among other psychological theories. LIDA’s implementation of consciousness enables us to account for the phenomenology of smooth coping, something that few cognitive architectures would be able to do. Through the fine granular analysis of LIDA, we argue that smooth coping is a sequence of automatized actions intermittently interspersed with consciously mediated action selection, supplemented by dorsal stream processes. In other words, non-conscious, automatized actions often require occasional bursts of conscious cognition to achieve the skillful and flexible adjustments of smooth coping. In addition, never-conscious dorsal stream information and associated sensorimotor processes provide further online adjustments during smooth coping. To achieve smooth coping in LIDA we introduce a new module to the LIDA cognitive architecture the Automatized Action Selection sub-module. Our complex model of smooth coping borrows notions of “embodied intelligence” from enactivism and augments these by allowing representations and more detailed mechanisms of conscious control. We explore several extended examples of smooth coping, starting from basic activities like walking and scaling up to more complex tasks like driving and short-order cooking. (shrink)

In this paper, I explore some of the ways in which Artificial Intelligence (AI) is mediated discursively. I assume that AI is informed by an “ancestral dream” to reproduce nature by artificial means. This dream drives the production of “cyborg discourse”, which hinges on the belief that human nature (especially intelligence) can be reduced to symbol manipulation and hence replicated in a machine. Cyborg discourse, I suggest, produces AI systems by rhetorical means; it does not merely describe AI systems or (...) reflect a set of prevailing attitudes about technology. To support this argument, I analyse a set of research articles about an “embodied conversational agent” called the Real Estate Agent (REA). The articles about REA mobilise a set of rhetorical strategies that systematically downplay the system’s artificiality and bolster its humanlike qualities. Within the context of the dream of AI to produce humanlike machines, and given our strong bias for human-human interaction, the designers’ claim to REA’s humanness in their research articles, as I argue in the final section of this paper, needs little justification. (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)

An examination of John Pollock's theory of artificial intelligence and philosophy of mind raises difficulties for his mechanist concept of person. Token physicalism, agent materialism, and strong artificial intelligence are so related that if the first two propositions are not well‐established, then there is no justification for believing that an artificial consciousness can be designed and built. Pollock's arguments are shown to be inconclusive in upholding a functionalist theory of persons as supervenient but purely physical entities. In part this is (...) the result of Pollock's thin definition of the concept of supervenience, according to which any complex supervenes on its proper parts. The limitations of this account are apparent when contrasted with richer conceptions of supervenience, such as Joseph Margolis’. But on Margolis’ theory, the mind and its expressions supervene on or rise above their material embodiments in the sense that they cannot be fully explained in physical terms, which contradicts Pollock's token physicalism and agent materialism. The consequence for Pollock's project to explain the mind as mechanical, and to manufacture artificial persons, is that these systems can at best aspire to impressive innovations in weak artificial intelligence, but realistically cannot aspire to strong or mentalistic artificial intelligence. (shrink)

Sedlakova and Trachsel (2023) consider conversational artificial intelligence (CAI) as a new way of providing psychotherapy to patients. This is an important topic, and Sedlakova and Trachsel have...

This article examines two medieval thinkers—Averroes and Aquinas—on the kind of causation exercised by the agent intellect in “abstracting” or producing intelligibles from images in the imagination. It argues that abstraction in these thinkers should be interpreted in causal terms, as an act whereby images in the imagination, through the power of the agent intellect, educe their intelligible likeness in a receptive intellect. This Averroan-Thomistic causal approach to abstraction offers an intriguing alternative to the usual approach to abstraction as an (...) epistemological content-sorting. The article also demonstrates the extensive common ground uniting these thinkers’ cognition theories, despite Aquinas’s well-known rejection of Averroes’s theory of separate Intellects. (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)