Ethics of Artificial Intelligence

This paper attempts to provide the basis for a broader naturalized account of agency. Naturalization is considered as the need for an ongoing and open-ended process of scientific inquiry driven by the continuous formulation of questions regarding a phenomenon. The naturalization of agency is focused around the interrelation of the fundamental notions of autonomy, functionality, intentionality and meaning. Certain naturalized frameworks of agency are critically considered in an attempt to bring together all the characteristic properties that constitute an autonomous agent, (...) as well as to indicate the shaping of these notions/properties during the development and the evolution of its agential capacity. Autonomy and interaction are proved to be key concepts in this endeavor. (shrink)

In my 2013 article, “A New Theory of Free Will”, I argued that several serious hypotheses in philosophy and modern physics jointly entail that our reality is structurally identical to a peer-to-peer (P2P) networked computer simulation. The present paper outlines how quantum phenomena emerge naturally from the computational structure of a P2P simulation. §1 explains the P2P Hypothesis. §2 then sketches how the structure of any P2P simulation realizes quantum superposition and wave-function collapse (§2.1.), quantum indeterminacy (§2.2.), wave-particle duality (§2.3.), (...) and quantum entanglement (§2.4.). Finally, §3 argues that although this is by no means a philosophical proof that our reality is a P2P simulation, it provides ample reasons to investigate the hypothesis further using the methods of computer science, physics, philosophy, and mathematics. (shrink)

The theory of mind debate has reached a “hybrid consensus” concerning the status of theory-theory and simulation-theory. Extant hybrid models either specify co-dependency and implementation relations, or distribute mentalizing tasks according to folk-psychological categories. By relying on a non-developmental framework these models fail to capture the central connection between simulation and theory. I propose a “dynamic” hybrid that is informed by recent work on the nature of similarity cognition. I claim that Gentner’s model of structure-mapping allows us to understand simulation (...) as a process in which psychological representations are aligned, causing the spontaneous abstraction of theoretical generalizations about the psychological domain. (shrink)

Mindreading is the ability to attribute mental states to other individuals. According to the Theory-Theory (TT), mindreading is based on one's possession of a Theory of Mind. On the other hand, the Simulation Theory (ST) maintains that one arrives at the attribution of a mental state by simulating it in one's own mind. In this paper, I propose a ST-TT hybrid model of the ability to attribute disgust on the basis of visual stimuli such as facial expressions, body postures, etc. (...) More precisely, while I defend Goldman's (2006) thesis that the ability to attribute disgust based on observing disgusted facial expressions stems from a mirror-based simulation process, I argue that ST is unable to account for the ability to attribute disgust based on non-facial visual stimuli; I propose, rather, that this latter ability is theory-based. My model is grounded in evidence from individuals suffering from Huntington's Disease. (shrink)

As we live in a culture where “everything can be commodified, measured and calculated and can be put in the competitive market for sale, detached from its roots and purpose,” there is need to redefine our humanness in terms of the changing nature of science, technology, and their deeper impact on human life. More than anything else, it is Information Technology that now has tremendous influence on all spheres of our life, and in a sense, IT has become the destiny (...) of our life. And this is where the real trauma lies. On the one hand, our being in the cyberspace opens up new and exciting horizons before us; on the other hand, we ourselves are changed and transformed in the process. The virtual world transforms human users to a problem-solver technocrat. The speed at which Information Technology is changing the way that youth around the world are socializing, playing, and researching, it is the common practice now for a 15-year old to go home and update their MySpace page, followed by playing online games, or looking up the new trendy YouTube video. These forms of technology are often the topic of adolescent conversations as YouTube, blogs, e-magazines, Face book, MySpace, iPhones, and iPods dominate the commercial and social networking market. Some researchers refer to this phenomenon as ubiquitous technology drawing attention to the fact that ubiquitous technology acknowledges the speedy adoption of day-to-day use of technology as a global phenomenon. In this background, this article aims at revisiting the question, “What is to be human in the era of Ubiquitous Technology?” From a feminist perspective, one can still redefine the boundaries between femininity and masculinity in the context of IT and its impact on our lifestyle and thought style. While examining the ways in which our definitions of “woman” and “man” are shifting in this new communication environment, Elizabeth Lane Lawley observes that we cannot fix a single center from which the experiences of women with computer and communication systems can be viewed and that such fixity would only serve to deepen inequities rather than exposing and removing them. She finally submits, “It is possible to use new theoretical perspectives on the shifting boundaries of gender definitions to rethink a previously deterministic view of the effect of new technologies on society, and particularly the effect of those technologies on women. While the gradual absence of the subject from the field of Artificial Intelligence leads to the invisibility of feminine care along with social and relational nature of man, some feminists dismiss the biological sex distinction on such issues and encourage females to ‘imitate man’ and to become more aggressive, assertive and dominating” (Lawley 1993 ). What are the possible impacts of this new technology on the so-called feminine traits of our human nature? How far our definitions of “woman” and “man” are shifting in this new communication environment? This is what this article seeks to explore. (shrink)

Representation was invented as an issue during the 17th century in response to specific developments in the technology of simulation. It remains an issue of central importance today in the design of information systems and approaches to artificial intelligence. Our cultural legacy of thought about representation is enormous but as inhibiting as it is productive. The challenge to designers of representative technology is to reshape this legacy by enlarging the politics rather than the technics of simulation.

_Intelligence Unbound_ explores the prospects, promises, and potential dangers of machine intelligence and uploaded minds in a collection of state-of-the-art essays from internationally recognized philosophers, AI researchers, science fiction authors, and theorists. Compelling and intellectually sophisticated exploration of the latest thinking on Artificial Intelligence and machine minds Features contributions from an international cast of philosophers, Artificial Intelligence researchers, science fiction authors, and more Offers current, diverse perspectives on machine intelligence and uploaded minds, emerging topics of tremendous interest Illuminates the nature (...) and ethics of tomorrow’s machine minds—and of the convergence of humans and machines—to consider the pros and cons of a variety of intriguing possibilities Considers classic philosophical puzzles as well as the latest topics debated by scholars Covers a wide range of viewpoints and arguments regarding the prospects of uploading and machine intelligence, including proponents and skeptics, pros and cons. (shrink)

Transhumanism is a loosely defined movement that has developed gradually over the past two decades. [1] It promotes an interdisciplinary approach to understanding and evaluating the opportunities for enhancing the human condition and the human organism opened up by the advancement of technology. Attention is given to both present technologies, like genetic engineering and information technology, and anticipated future ones, such as molecular nanotechnology and artificial intelligence.

Societal pressures, accreditation organizations, and licensing agencies are emphasizing the importance of ethics in the engineering curriculum. Traditionally, this subject has been taught using dogma, heuristics, and case study approaches. Most recently a number of organizations have sought to increase the utility of these approaches by utilizing the Internet. Resources from these organizations include on-line courses and tests, videos, and DVDs. While these individual approaches provide a foundation on which to base engineering ethics, they may be limited in developing a (...) student’s ability to identify, analyze, and respond to engineering ethics situations outside of the classroom environment. More effective approaches utilize a combination of these types of approaches. This paper describes the design and development of an internet based interactive Simulator for Engineering Ethics Education. The simulator places students in first person perspective scenarios involving different types of ethical situations. Students must gather data, assess the situation, and make decisions. This requires students to develop their own ability to identify and respond to ethical engineering situations. A limited comparison between the internet based interactive simulator and conventional internet web based instruction indicates a statistically significant improvement of 32% in instructional effectiveness. The simulator is currently being used at the University of Houston to help fulfill ABET requirements. (shrink)

Viewed from an evolutionary standpoint, emotions can be understood as situation-specific patterns of energy consumption related to behaviors that have been selected by evolution for their survival value, such as environmental exploration, flight or fight, and socialization. In the present article, the energy linked with emotions is investigated by a strictly energy-based simulation of the evolution of simple autonomous agents provided with random cognitive and motor capacities and operating among food and predators. Emotions are translated into evolving patterns of energy (...) consumption related to situation-specific behaviors. As a result, a variety of behaviors resembling emotions emerge, each with a specific pattern of energy consumption. There is little difference between emotion-like behaviors on individual and collective levels. The fact that these patterns evolve under a strictly energetic selection regime indicates the interesting role and adaptive advantage of emotions as situation-specific energy savers. Although gained through an approach characterized by a highly reduced complexity, these findings support a quantitative understanding of emotions that may improve the existing qualitative approaches and could have far-reaching implications for further research on and conceptualization of emotions and their interactions with cognition. (shrink)

Andy Clark is currently Professor of Philosophy and Director of the Philosophy/Neuroscience/Psychology program at Washington University in St. Louis, Missouri. He is the author of two books MICROCOGNITION (MIT Press/Bradford Books 1989) and ASSOCIATIVE ENGINES (MIT Press/Bradford Books, 1993) as well as numerous papers and four edited volumes. He is an ex- committee member of the British Society for the Philosophy of Science and of the Society for Artificial Intelligence and the Simulation of Behavior. Awards include a visiting Fellowship at (...) the Australia National University, Canberra in 1989, and an ESRC Senior Research Leave Fellowship in 1992. He is an Associate of BEHAVIORAL AND BRAIN SCIENCES, serves on the editorial boards of the journals. (shrink)

The aim of this project was to explore the possibility of constructing a psychologic inference engine that might enhance introspective self-awareness by delivering inferences about a user based on what he said in interactive dialogues about his closest opposite-sex relation. To implement this aim, we developed a computer program (guru) with the capacity to simulate human conversation in colloquial natural language. The psychologic inferences offered represent the authors' simulations of their commonsense psychology responses to expected user-input expressions. The heuristics of (...) the natural language processor and its relation to output responses are described in enough detail for the operations of the implementation to be understood. Evaluation of this new cognitive agent presents, we hope, puzzles for artificial intelligence and cognitive science. (shrink)

Modern epistemic questions have largely been focused around the individual and her ability to acquire knowledge autonomously. More recently epistemologists have begun to look more broadly in providing accounts of knowledge by considering its social context, where the individual depends on others for true beliefs. Hardwig explains the effect of this shift starkly, arguing that to reject epistemic dependency is to deny certain true beliefs widely held throughout society and, more specifically, it is to deny that science and scholarship can (...) provide true belief. Alternatively, Hardwig argues that beliefs could be granted to communities or groups but denied to individuals. This paper approaches these broad assertions using a group agency model from List and Pettit. Through a discussion of the of group agents, I conclude that List and Pettit give us reason to accept some of Hardwig's concerns, but that attributing beliefs to groups does not require us to deny them to individuals, rather an individual can use a group agent as a source of epistemic dependence. (shrink)

CHAPTER Simulation theory and mental concepts Alvin I. Goldman Rutgers University. Folk psychology and the TT-ST debate The study of folk psychology, ...

of autonomous agency. Yet neither denies the importance of human freedom. In Heidegger's early work the subject is reinterpreted as Dasein -- a non autonomous, culturally bound (or thrown) way of being, that can yet change the field of possibilities in which it acts. In middle Heidegger, thinkers alone have the power to disclose a new world, while in later Heidegger, anyone is free to step back from the current world, to enter one of a plurality of worlds, and, thereby, (...) facilitate a change in the practices of one's society. Likewise, for early Foucault, the subject is reduced to a function of discourse; for middle Foucault, writing can open up new worlds, and in later Foucault, freedom is understood as the power to question what is currently taken for granted, plus the capacity to change oneself and, perhaps, one's milieu. In short, while both Heidegger and Foucault reject the Enlightenment idea of an autonomous subject, they have a robust notion of freedom and action. And it will turn out for both thinkers that each person can modify his or her cultural practices by openness to embeddedness in them. All this needs a great deal of explanation. We need to determine, on the one hand, just what each rejects and why, and, on the other, what series of understandings of the self and its possibilities for action each introduces. (shrink)

Accessing legislation via the Internet is more and more frequent. As a result, systems that allow consultation of law texts are becoming more and more powerful. This paper presents DARES, a generic system which can be adapted to any domain to handle documents production needs. It is based on an annotation engine which allows obtaining XML documents inputs as required by the system, and on an XML fragments recombining system. The latter operates using a fragment manipulation functions toolbox to generate (...) new documents. To validate this system, we have tried to apply it to the domain of law through the consolidation problem. (shrink)

The assessment of highly domain-general problem solving skills is increasingly important as problem solving is increasingly demanded by modern workplaces and increasingly present in international large-scale assessments such as the Programme for International Student Assessment. This thesis is about the computer-based assessment of problem solving skills based on Multiple Complex Systems : The main idea of the MCS approach is to present multiple computer-simulations of “minimally complex” problems in order to reliably assess certain problem solving skills. In each simulation, the (...) problem solver has to interact with a problem in order to find out how to adequately represent the problem, and how to solve the problem. Up to now, two instances of the MCS approach have been proposed: the MicroDYN approach and – more recently, in the second paper of this thesis – the MicroFIN approach. In the current thesis I will elaborate on three research questions regarding the validity of the MCS approach: its content validity with regard to the concept of complex problem solving; the convergent validity of different instances of the MCS approach; the discriminant validity of the interactive problems of the MCS approach with regard to traditional static measures of reasoning and analytic problem solving skills. Each research question will be addressed in one corresponding paper: In a first paper complex problem solving is defined as the goal-oriented control of systems that contain multiple highly interrelated elements. After reviewing some of the major strands of research on complex problem solving a theoretical framework outlining the most important cognitive processes involved in solving complex problems is derived. The theoretical framework highlights both interactive knowledge acquisition and interactive knowledge application as the two major phases in the process of complex problem solving. Both phases are represented in all current instances of the MCS approach. In a second paper the convergent validity of MicroDYN and MicroFIN is investigated in order to demonstrate that both instances address the same kind of problem solving skills. Based on a multitrait-multimethod analysis of a sample of university students it is demonstrated that – in addition to method-specific skills – both instances assess a common set of skills related to representing and solving different kinds of interactive problems. In a regression of science grades on reasoning and the skills assessed by the instances of the MCS approach it is demonstrated that only the method-general representation trait and reasoning have substantial unique contributions. Thus, MicroDYN and MicroFIN seem to address a common set of skills and this set of skills is relevant for explaining school grades in science classes even beyond reasoning. In a third paper the discriminant validity of the interactive MicroDYN test is investigated by relating it to reasoning and traditional static measures of Analytic Problem Solving skills as they were applied in PISA 2003. Besides a common core of problem solving skills addressed by both kinds of tasks Fischer et al. expected to find evidence for additional skills that were related to interactive problems only. Results indicate that MicroDYN shares a lot of variance with APS even after controlling for reasoning in a sample of high-school students and the university student sample. With regard to the explanation of school grades MicroDYN had an incremental value compared to reasoning and APS in the high-school student sample but not significantly so in the university student sample. Basically these findings highlight both potential and limitations of the MicroDYN approach in its current form. Current instances of the MCS approach address a small set of problem solving skills reliably, but it takes more than these skills to competently solve complex problems. Implications for future research on the assessment of problem solving skills are discussed. (shrink)

This paper contributes to an ongoing debate regarding the cognitive processes involved when one person predicts a target person's behavior and/or attributes a mental state to that target person. According to simulation theory, a person typically performs these tasks by employing some part of her brain as a simulation of what is going on in a corresponding part of the brain of the target person. I propose a general intuitive analysis of what 'simulation' means. Simulation is a particular way of (...) using one process to acquire knowledge about another process. What distinguishes simulation from other ways of acquiring knowledge is that simulation requires, for its non-accidental success, that the simulating process reflect significant aspects of the simulated process. This conceptual work is of independent philosophical interest, but it also enables me to argue for two conclusions that are of great significance to the debate about mental simulation theory. First, I argue that, in order to stake a non-trivial claim, simulation theory must hold that mental simulation involves what I call concretely similar processes. Second, I argue for the surprising conclusion that a significant class of cases that simulation theorists have claimed as intuitive cases of simulation do not actually involve simulation, after all. I close by sketching an alternative account that might handle these problematic cases. (shrink)

The philosophy of information (PI) is a new area of research with its own field of investigation and methodology. This article, based on the Herbert A. Simon Lecture of Computing and Philosophy I gave at Carnegie Mellon University in 2001, analyses the eighteen principal open problems in PI. Section 1 introduces the analysis by outlining Herbert Simon's approach to PI. Section 2 discusses some methodological considerations about what counts as a good philosophical problem. The discussion centers on Hilbert's famous analysis (...) of the central problems in mathematics. The rest of the article is devoted to the eighteen problems. These are organized into five sections: problems in the analysis of the concept of information, in semantics, in the study of intelligence, in the relation between information and nature, and in the investigation of values. (shrink)

This paper supports the view that the ongoing shift from orthodox to embodied-embedded cognitive science has been significantly influenced by the experimental results generated by AI research. Recently, there has also been a noticeable shift toward enactivism, a paradigm which radicalizes the embodied-embedded approach by placing autonomous agency and lived subjectivity at the heart of cognitive science. Some first steps toward a clarification of the relationship of AI to this further shift are outlined. It is concluded that the success of (...) enactivism in establishing itself as a mainstream cognitive science research program will depend less on progress made in AI research and more on the development of a phenomenological pragmatics. (shrink)

from non-conscious components by positing that consciousness is a universal primitive. For example, the double aspect theory of information holds that infor- mation has a phenomenal aspect. How then do you get from phenomenal infor- mation to human consciousness? This paper proposes that an entity is conscious to the extent it amplifies information, first by trapping and integrating it through closure, and second by maintaining dynamics at the edge of chaos through simul- taneous processes of divergence and convergence. The origin (...) of life through autocatalytic closure, and the origin of an interconnected worldview through conceptual closure, induced phase transitions in the degree to which informa- tion, and thus consciousness, is locally amplified. Divergence and convergence of cognitive information may involve phenomena observed in light e.g. focusing, interference, and resonance. By making information flow inward- biased, clo- sure shields us from external consciousness; thus the paucity of consciousness may be an illusion. (shrink)

In this paper, I hope to show how a recent theory in the philosophy of mind concerning how we ‘read’ the minds of others – namely, Heal’s version of simulation theory – is consistent with the view that the kind of understanding we bring to bear on the irrational is different in kind from the way we understand one another in the course of everyday life. I shall attempt to show that Heal’s version of simulation theory (co-cognition) is to be (...) favoured over its rival ‘theory theories’ in the light of its accommodation of deeply irrational mental states. I claim that simulation theory preserves an asymmetry of psychological explanation which is not similarly preserved by a theory theory account, and I argue that this preservation of asymmetry is an advantage of Heal’s account. (shrink)

The same neural structures involved in the unconscious modeling of our acting body in space also contribute to our awareness of the lived body and of the objects that the world contains. Neuroscientific research also shows that there are neural mechanisms mediating between the multi-level personal experience we entertain of our lived body, and the implicit certainties we simultaneously hold about others. Such personal and body-related experiential knowledge enables us to understand the actions performed by others, and to directly decode (...) the emotions and sensations they experience. A common functional mechanism is at the basis of both body awareness and basic forms of social understanding: embodied simulation. It will be shown that the present proposal is consistent with some of the perspectives offered by phenomenology. (shrink)

de Bruin & Gallagher suggest that the view of embodied simulation put forward in our recent article lacks explanatory power. We argue that the notion of reuse of mental states represented with a bodily format provides a convincing simulational account of the mirroring mechanism and its role in mind -reading.

Simulation theories of social cognition abound in the literature, but it is often unclear what simulation means and how it works. The discovery of mirror neurons, responding both to action execution and observation, suggested an embodied approach to mental simulation. Over the last years this approach has been hotly debated and alternative accounts have been proposed. We discuss these accounts and argue that they fail to capture the uniqueness of embodied simulation (ES). ES theory provides a unitary account of basic (...) social cognition, demonstrating that people e their own mental states or processes represented with a bodily format in functionally attributing them to others. (shrink)

There has been considerable discussion in the past about the assumptions and basis of different ethical rules. For instance, it is commonplace to say that ethical rules are defaults rules, which means that they tolerate exceptions. Some authors argue that morality can only be grounded in particular cases while others defend the existence of general principles related to ethical rules. Our purpose here is not to justify either position, but to try to model general ethical rules with artificial intelligence formalisms (...) and to compute logical consequences of different ethical theories. More precisely, this is an attempt to show that progress in non-monotonic logics, which simulates default reasoning, could provide a way to formalize different ethical conceptions. From a technical point of view, the model developed in this paper makes use of the Answer Set Programming (ASP) formalism. It is applied comparatively to different ethical systems with respect to their attitude towards lying. The advantages of such formalization are two-fold: firstly, to clarify ideas and assumptions, and, secondly, to use solvers to derive consequences of different ethical conceptions automatically, which can help in a rigorous comparison of ethical theories. (shrink)

A computational theory of induction must be able to identify the projectible predicates, that is to distinguish between which predicates can be used in inductive inferences and which cannot. The problems of projectibility are introduced by reviewing some of the stumbling blocks for the theory of induction that was developed by the logical empiricists. My diagnosis of these problems is that the traditional theory of induction, which started from a given (observational) language in relation to which all inductive rules are (...) formulated, does not go deep enough in representing the kind of information used in inductive inferences. As an interlude, I argue that the problem of induction, like so many other problems within AI, is a problem of knowledge representation. To the extent that AI-systems are based on linguistic representations of knowledge, these systems will face basically the same problems as did the logical empiricists over induction. In a more constructive mode, I then outline a non-linguistic knowledge representation based on conceptual spaces. The fundamental units of these spaces are "quality dimensions". In relation to such a representation it is possible to define "natural" properties which can be used for inductive projections. I argue that this approach evades most of the traditional problems. (shrink)

Simulation has emerged as an increasingly popular account of folk psychological (FP) talents at mind-reading: predicting and explaining human mental states. Where its rival (the theory-theory) postulates that these abilities are explained by mastery of laws describing the connections between beliefs, desires, and action, simulation theory proposes that we mind-read by "putting ourselves in another's shoes." This paper concerns connectionist architecture and the debate between simulation theory (ST) and the theory-theory (TT). It is only natural to associate TT with classical (...) architectures where rule governed operations apply to explicit propositional representations. On the other hand, ST would seem better tuned to procedurally oriented non-symbolic structures found in connectionist models. This paper explores the possible alignment between ST and connectionist architecture. Joe Cruz argues that connectionist models with distributed non-symbolic representations are particularly well suited to simulation theory. The purported linkage between connectionist architecture and simulation theory is criticized in this paper. The conclusion is that there are reasons for thinking that connectionist forms of representation are the enemy of both TT and ST. So the contribution of connectionism may be to suggest the need for an alternative to both views. (shrink)

This paper considers the impact of the AI R&D programme on human society and the individual human being on the assumption that a full realisation of the engineering objective of AI, namely, construction of human-level, domain-independent intelligent entities, is possible. Our assumption is essentially identical tothe maximum progress scenario of the Office of Technology Assessment, US Congress.Specifically, the first section introduces some of the significant issues on the relational nexus among work, education and the human-machine boundary. In particular, based on (...) a Russellian conception of rationality I briefly argue that we need to change our related conceptions of work, employment and free time, through a new human-centred education. On the human-machine boundary problem, I make a couple of tentative suggestions and put forward some crucial open questions.Section two discusses the impact of the emerging machine intelligence on human nature both as modification of its self-image, keeping human nature itself unchanged, and its potential for altering human nature itself. I briefly argue that: (i) in a certain context, the question of the supremacy or uniqueness of human intelligence loses much, if not all, of its ‘weight’; and (ii) appearance of Robot-X species would immortalise the human spirit. (shrink)

According to the "experimenter's regress", disputes about the validity of experimental results cannot be closed by objective facts because no conclusive criteria other than the outcome of the experiment itself exist for deciding whether the experimental apparatus was functioning properly or not. Given the frequent characterization of simulations as "computer experiments", one might worry that an analogous regress arises for computer simulations. The present paper analyzes the most likely scenarios where one might expect such a "simulationist's regress" to surface, and, (...) in doing so, discusses analogies and disanalogies between simulation and experimentation. I conclude that, on a properly broadened understanding of robustness, the practice of simulating mathematical models can be seen to have sufficient internal structure to avoid any special susceptibility to regress-like situations. (shrink)

Apocalyptic AI, the hope that we might one day upload our minds into machines and live forever in cyberspace, has become commonplace. This view now affects robotics and AI funding, play in online games, and philosophical and theological conversations about morality and human dignity.

In considering how to best deploy robotic systems in public and private sectors, we must consider what individuals will expect from the robots with which they interact. Public awareness of robotics—as both military machines and domestic helpers—emerges out of a braided stream composed of science fiction and popular science. These two genres influence news media, government and corporate spending, and public expectations. In the Euro-American West, both science fiction and popular science are ambivalent about the military applications for robotics, and (...) thus we can expect their readers to fear the dangers posed by advanced robotics while still eagerly anticipating the benefits to be accrued through them. The chief pop science authors in robotics and artificial intelligence have a decidedly apocalyptic bent and have thus been described as leaders in a social movement called "Apocalyptic AI." In one form or another, such authors look forward to a transcendent future in which machine life succeeds human life, thanks to the march of evolutionary progress. The apocalyptic promises of popular robotics presume that presently exponential growth in computing will continue indefinitely, producing a "Singularity." During the Singularity, technological progress will be so rapid that undreamt of changes will take place on earth, the most important of which will be the evolutionary succession of human beings by massively intelligent robots and the "uploading" of human consciousness into computer bodies. This supposedly inevitable transition into post-biological life looms across the entire scope of pop robotics and artificial intelligence, and it is from beneath that shadow that all popular books engage the military and the ethics of warfare. Creating a just future will require that we transcend the apocalyptic discourse of pop science and establish an ethical approach to researching and deploying robots, one that emphasizes human rather than robot welfare; doing so will require the collaboration of social scientists, humanists, and scientists. (shrink)

Morrison points out many similarities between the roles of simulation models and other sorts of models in science. On the basis of these similarities she claims that running a simulation is epistemologically on a par with doing a traditional experiment and that the output of a simulation therefore counts as a measurement. I agree with her premises but reject the inference. The epistemological payoff of a traditional experiment is greater (or less) confidence in the fit between a model and a (...) target system. The source of this payoff is the existence of a causal interaction with the target system. A computer experiment, which does not go beyond the simulation system itself, lacks any such interaction. So computer experiments cannot confer any additional confidence in the fit (or lack thereof) between the simulation model and the target system. (shrink)

Socio-ethics covers the relation of the individual with the group and with society, as the individual acquires the skills for social life with others and the conduct of ‘normal responsible behaviour’ (Leal in AI Soc 9:29–32, 1995) that guides moral action. For a consideration of what it means to be socially skilled in everyday human interaction and the ethical issues arising from the new conditions of interaction that come with the integration of intelligent interactive artefacts, we will provide an analysis (...) at multiple levels of these phenomena and draw on a variety of application domains, for example, healthcare and interactive media. (shrink)

Neuroscience and technological medicine in general increasingly faces us with the imminent reality of cyborgs—integrated part human and part machine complexes.If my brain functions in a way that is supported by and exploits intelligent technology both external and implantable, then how should I be treated and what is my moral status—am I a machine or am I a person? I explore a number of scenarios where the balance between human and humanoid machine shifts, and ask questions about the moral status (...) of the individuals concerned. The position taken is very much in accordance with the Aristotelian idea that our moral behaviour is of a piece with our social and personal skills and forms a reactive and reflective component of those skills. (shrink)

Much has been written about the possibility of human trust in robots. In this article we consider a more specific relationship: that of a human follower’s obedience to a social robot who leads through the exercise of referent power and what Weber described as ‘charismatic authority.’ By studying robotic design efforts and literary depictions of robots, we suggest that human beings are striving to create charismatic robot leaders that will either (1) inspire us through their display of superior morality; (2) (...) enthrall us through their possession of superhuman knowledge; or (3) seduce us with their romantic allure. Rejecting a contractarian-individualist approach which presumes that human beings will be able to consciously ‘choose’ particular robot leaders, we build on the phenomenological-social approach to trust in robots to argue that charismatic robot leaders will emerge naturally from our world’s social fabric, without any rational decision on our part. Finally, we argue that the stability of these leader-follower relations will hinge on a fundamental, unresolved question of robotic intelligence: is it possible for synthetic intelligences to exist that are morally, intellectually, and emotionally sophisticated enough to exercise charismatic authority over human beings—but not so sophisticated that they lose the desire to do so? (shrink)

While it seems unlikely that any method of guaranteeing human-friendliness on the part of advanced Artificial General Intelligence systems will be possible, this doesn’t mean the only alternatives are throttling AGI development to safeguard humanity, or plunging recklessly into the complete unknown. Without denying the presence of a certain irreducible uncertainty in such matters, it is still sensible to explore ways of biasing the odds in a favorable way, such that newly created AI systems are significantly more likely than not (...) to be Friendly. Several potential methods of effecting such biasing are explored here, with a particular but non-exclusive focus on those that are relevant to open-source AGI projects, and with illustrative examples drawn from the OpenCog open-source AGI project. Issues regarding the relative safety of open versus closed approaches to AGI are discussed and then nine techniques for biasing AGIs in favor of Friendliness are presented: 1. Engineer the capability to acquire integrated ethical knowledge.2. Provide rich ethical interaction and instruction, respecting developmental stages.3. Develop stable, hierarchical goal systems.4. Ensure that the early stages of recursive self-improvement occur relatively slowly and with rich human involvement.5. Tightly link AGI with the Global Brain.6. Foster deep, consensus-building interactions between divergent viewpoints.7. Create a mutually supportive community of AGIs.8. Encourage measured co-advancement of AGI software and AGI ethics theory.9. Develop advanced AGI sooner not later. In conclusion, and related to the final point, we advise the serious co-evolution of functional AGI systems and AGI-related ethical theory as soon as possible, before we have so much technical infrastructure that parties relatively unconcerned with ethics are able to rush ahead with brute force approaches to AGI development. (shrink)

We introduce an Automatic Theorem Prover (ATP) of a dual tableau system for a relational logic for order of magnitude qualitative reasoning, which allows us to deal with relations such as negligibility, non-closeness and distance. Dual tableau systems are validity checkers that can serve as a tool for verification of a variety of tasks in order of magnitude reasoning, such as the use of qualitative sum of some classes of numbers. In the design of our ATP, we have introduced some (...) heuristics, such as the so called phantom variables, which improve the efficiency of the selection of variables used un the proof. (shrink)

The boundaries of the ethical have traditionally coincided with the boundaries of humanity. This, however, is no longer the case. Scientific developments, such as genetic engineering, stem-cell research, cloning, the Human Genome Project, new paleontological evidence, and the rise of neuropsychology call into question the very notion of human being and thus require a new conceptual map for ethical judgment. The contours of this map may be seen to emerge in works of science fiction (SF), which not only vividly dramatize (...) the implications and consequences of new technologies and discoveries, but also exert a powerful influence on culture, creating a feedback loop of images and ideas. This essay focuses on three SF topoi: the human/animal evolutionary boundary; non-biological subjectivity (AI); and the human/alien interaction. It explores each of these topoi in a selection of SF texts, including novels by H. G. Wells, Olaf Stapledon, Stephen Baxter, William Gibson, Stanislaw Lem and others, showing how the boundaries of humanity are expanded and then exploded through the radical subversion of the tenets of liberal humanism. (shrink)

In their important paper “Autonomous Agents”, Floridi and Sanders use “levels of abstraction” to argue that computers are or may soon be moral agents. In this paper we use the same levels of abstraction to illuminate differences between human moral agents and computers. In their paper, Floridi and Sanders contributed definitions of autonomy, moral accountability and responsibility, but they have not explored deeply some essential questions that need to be answered by computer scientists who design artificial agents. One such question (...) is, “Can an artificial agent that changes its own programming become so autonomous that the original designer is no longer responsible for the behavior of the artificial agent?” To explore this question, we distinguish between LoA1 (the user view) and LoA2 (the designer view) by exploring the concepts of unmodifiable, modifiable and fully modifiable tables that control artificial agents. We demonstrate that an agent with an unmodifiable table, when viewed at LoA2, distinguishes an artificial agent from a human one. This distinction supports our first counter-claim to Floridi and Sanders, namely, that such an agent is not a moral agent, and the designer bears full responsibility for its behavior. We also demonstrate that even if there is an artificial agent with a fully modifiable table capable of learning* and intentionality* that meets the conditions set by Floridi and Sanders for ascribing moral agency to an artificial agent, the designer retains strong moral responsibility. (shrink)

The field of Artificial Intelligence has been around for over 60 years now. Soon after its inception, the founding fathers predicted that within a few years an intelligent machine would be built. That prediction failed miserably. Not only hasn’t an intelligent machine been built, but we are not much closer to building one than we were some 50 years ago. Many reasons have been given for this failure, but one theme has been dominant since its advent in 1969: The Frame (...) Problem. What looked initially like an innocuous problem in logic, turned out to be a much broader and harder problem of holism and relevance in commonsense reasoning. Despite an enormous literature on the topic, there is still disagreement not only on whether the problem has been solved, but even what exactly the problem is. In this paper we provide a formal description of the initial problem, the early attempts at a solution, and its ramification both in AI as well as philosophy. (shrink)