New AI technologies have the potential to cause unintended harms in diverse domains including warfare, judicial sentencing, biomedicine and governance. One strategy for realising the benefits of AI whilst avoiding its potential dangers is to ensure that new AIs are properly ‘aligned’ with some form of ‘alignment target.’ One danger of this strategy is that – dependent on the alignment target chosen – our AIs might optimise for objectives that reflect the values only of a certain subset of society, and (...) that do not take into account alternative views about what constitutes desirable and safe behaviour for AI agents. In response to this problem, several AI ethicists have suggested alignment targets that are designed to be sensitive to widespread normative disagreement amongst the relevant stakeholders. Authors inspired by voting theory have suggested that AIs should be aligned with the verdicts of actual or simulated ‘moral parliaments’ whose members represent the normative views of the relevant stakeholders. Other authors inspired by decision theory and the philosophical literature on moral uncertainty have suggested that AIs should maximise socially expected choiceworthiness. In this paper, I argue that both of these proposals face several important problems. In particular, they fail to select attractive ‘compromise options’ in cases where such options are available. I go on to propose and defend an alternative, bargaining-theoretic alignment target, which avoids the problems associated with the voting- and decision-theoretic approaches. (shrink)

The rapid progress of artificial intelligence (AI) has resulted in its integration into various stages of the media process, including information gathering, processing, and distribution. This integration has raised the possibility of AI dominating the media industry, leading to an era of “autonomous driving” within AI-driven media systems. Similar to the ethical dilemma known as the “trolley problem” (TP) in autonomous driving, a comparable problem arises in AI automated media. This study examines the emergence of the new TP in fully (...) automated AI-driven media (FAAIM), recognizing the complexity of this problem given the nature of media content and its societal impact. To address this complexity, we propose the adoption of the theory of “meaningful human control,” originally developed to address responsibility in the governance of lethal autonomous weapons systems, as a framework for governing FAAIM. By ensuring that humans can be held responsible for the operations of FAAIM, this paper aims to proactively confront the ethical challenges arising from FAAIM, identify potential solutions, and guide future research in media ethics. (shrink)

Language agents are AI systems capable of understanding and responding to natural language, potentially facilitating the process of encoding human goals into AI systems. However, this paper argues that if language agents can achieve easy alignment, they also increase the risk of malevolent agents building harmful AI systems aligned with destructive intentions. The paper contends that if training AI becomes sufficiently easy or is perceived as such, it enables malicious actors, including rogue states, terrorists, and criminal organizations, to create powerful (...) AI systems devoted to their nefarious aims. Given the strong incentives for such groups and the rapid progress in AI capabilities, this risk demands serious attention. In addition, the paper highlights considerations suggesting that the negative impacts of language agents may outweigh the positive ones, including the potential irreversibility of certain negative AI impacts. The overarching lesson is that various AI-related issues are intimately connected with each other, and we must recognize this interconnected nature when addressing those issues. (shrink)

We argue that two long-term goals of AI research stand in tension with one another. The first involves creating AI that is safe, where this is understood as solving the problem of value alignment. The second involves creating artificial general intelligence, meaning AI that operates at or beyond human capacity across all or many intellectual domains. Our argument focuses on the human capacity to make what we call “existential choices”, choices that transform who we are as persons, including transforming what (...) we most deeply value or desire. It is a capacity for a kind of value misalignment, in that the values held prior to making such choices can be significantly different from (misaligned with) the values held after making them. Because of the connection to existentialist philosophers who highlight these choices, we call the resulting form of risk “existentialist risk.” It is, roughly, the risk that results from AI taking an active role in authoring its own values rather than passively going along with the values given to it. On our view, human-like intelligence requires a human-like capacity for value misalignment, which is in tension with the possibility of guaranteeing value alignment between AI and humans. (shrink)

Much has been written about the need for trustworthy artificial intelligence (AI), but the underlying meaning of trust and trustworthiness can vary or be used in confusing ways. It is not always clear whether individuals are speaking of a technology’s trustworthiness, a developer’s trustworthiness, or simply of gaining the trust of users by any means. In sociotechnical circles, trustworthiness is often used as a proxy for ‘the good’, illustrating the moral heights to which technologies and developers ought to aspire, at (...) times with a multitude of diverse requirements; or at other times, no specification at all. In philosophical circles, there is doubt that the concept of trust should be applied at all to technologies rather than their human creators. Nevertheless, people continue to intuitively reason about trust in technologies in their everyday language. This qualitative study employed an empirical ethics methodology to address how developers and users define and construct requirements for trust throughout development and use, through a series of interviews. We found that different accounts of trust (rational, affective, credentialist, norms based, relational) served as the basis for individual granting of trust in technologies and operators. Ultimately, the most significant requirement for user trust and assessment of trustworthiness was the accountability of AI developers for the outputs of AI systems, hinging on the identification of accountable moral agents and perceived value alignment between the user and developer’s interests. (shrink)

Why should moral philosophers, moral psychologists, and machine ethicists care about computational complexity? Debates on whether artificial intelligence (AI) can or should be used to solve problems in ethical domains have mainly been driven by what AI can or cannot do in terms of human capacities. In this paper, we tackle the problem from the other end by exploring what kind of moral machines are possible based on what computational systems can or cannot do. To do so, we analyze normative (...) ethics through the lens of computational complexity. First, we introduce computational complexity for the uninitiated reader and discuss how the complexity of ethical problems can be framed within Marr’s three levels of analysis. We then study a range of ethical problems based on consequentialism, deontology, and virtue ethics, with the aim of elucidating the complexity associated with the problems themselves (e.g., due to combinatorics, uncertainty, strategic dynamics), the computational methods employed (e.g., probability, logic, learning), and the available resources (e.g., time, knowledge, learning). The results indicate that most problems the normative frameworks pose lead to tractability issues in every category analyzed. Our investigation also provides several insights about the computational nature of normative ethics, including the differences between rule- and outcome-based moral strategies, and the implementation-variance with regard to moral resources. We then discuss the consequences complexity results have for the prospect of moral machines in virtue of the trade-off between optimality and efficiency. Finally, we elucidate how computational complexity can be used to inform both philosophical and cognitive-psychological research on human morality by advancing the moral tractability thesis. (shrink)

Responsible AI (RAI) guidelines aim to ensure that AI systems respect democratic values. While a step in the right direction, they currently fail to impact practice. Our work discusses reasons for this lack of impact and clusters them into five areas: (1) the abstract nature of RAI guidelines, (2) the problem of selecting and reconciling values, (3) the difficulty of operationalising RAI success metrics, (4) the fragmentation of the AI pipeline, and (5) the lack of internal advocacy and accountability. Afterwards, (...) we introduce a number of approaches to RAI from a range of disciplines, exploring their potential as solutions to the identified challenges. We anchor these solutions in practice through concrete examples, bridging the gap between the theoretical considerations of RAI and on-the-ground processes that currently shape how AI systems are built. Our work considers the socio-technical nature of RAI limitations and the resulting necessity of producing socio-technical solutions. (shrink)

Generative AI enables automated, effective manipulation at scale. Despite the growing general ethical discussion around generative AI, the specific manipulation risks remain inadequately investigated. This article outlines essential inquiries encompassing conceptual, empirical, and design dimensions of manipulation, pivotal for comprehending and curbing manipulation risks. By highlighting these questions, the article underscores the necessity of an appropriate conceptualisation of manipulation to ensure the responsible development of Generative AI technologies.

Recent progress in artificial intelligence (AI) has drawn attention to the technology’s transformative potential, including what some see as its prospects for causing large-scale harm. We review two influential arguments purporting to show how AI could pose catastrophic risks. The first argument — the Problem of Power-Seeking — claims that, under certain assumptions, advanced AI systems are likely to engage in dangerous power-seeking behavior in pursuit of their goals. We review reasons for thinking that AI systems might seek power, that (...) they might obtain it, that this could lead to catastrophe, and that we might build and deploy such systems anyway. The second argument claims that the development of human-level AI will unlock rapid further progress, culminating in AI systems far more capable than any human — this is the Singularity Hypothesis. Power-seeking behavior on the part of such systems might be particularly dangerous. We discuss a variety of objections to both arguments and conclude by assessing the state of the debate. (shrink)

When making substituted judgments for incapacitated patients, surrogates often struggle to guess what the patient would want if they had capacity. Surrogates may also agonize over having the (sole) responsibility of making such a determination. To address such concerns, a Patient Preference Predictor (PPP) has been proposed that would use an algorithm to infer the treatment preferences of individual patients from population-level data about the known preferences of people with similar demographic characteristics. However, critics have suggested that even if such (...) a PPP were more accurate, on average, than human surrogates in identifying patient preferences, the proposed algorithm would nevertheless fail to respect the patient’s (former) autonomy since it draws on the ‘wrong’ kind of data: namely, data that are not specific to the individual patient and which therefore may not reflect their actual values, or their reasons for having the preferences they do. Taking such criticisms on board, we here propose a new approach: the Personalized Patient Preference Predictor (P4). The P4 is based on recent advances in machine learning, which allow technologies including large language models to be more cheaply and efficiently ‘fine-tuned’ on person-specific data. The P4, unlike the PPP, would be able to infer an individual patient’s preferences from material (e.g., prior treatment decisions) that is in fact specific to them. Thus, we argue, in addition to being potentially more accurate at the individual level than the previously proposed PPP, the predictions of a P4 would also more directly reflect each patient’s own reasons and values. In this article, we review recent discoveries in artificial intelligence research that suggest a P4 is technically feasible, and argue that, if it is developed and appropriately deployed, it should assuage some of the main autonomy-based concerns of critics of the original PPP. We then consider various objections to our proposal and offer some tentative replies. (shrink)

Ambitious value learning proposals to solve the AI alignment problem and avoid catastrophic outcomes from a possible future misaligned artificial superintelligence (such as Coherent Extrapolated Volition [CEV]) have focused on ensuring that an artificial superintelligence (ASI) would try to do what humans would want it to do. However, present and future sentient non-humans, such as non-human animals and possible future digital minds could also be affected by the ASI’s behaviour in morally relevant ways. This paper puts forward Sentientist Coherent Extrapolated (...) Volition, an alternative to CEV, that directly takes into account the interests of all sentient beings. This ambitious value learning proposal would significantly reduce the likelihood of risks of astronomical suffering from the ASI’s behaviour, and thus we have very strong pro-tanto moral reasons in favour of implementing it instead of CEV. This fact is crucial in conducting an adequate cost-benefit analysis between different ambitious value learning proposals. (shrink)

Artificial Intelligence (AI) is increasingly adopted in society, creating numerous opportunities but at the same time posing ethical challenges. Many of these are familiar, such as issues of fairness, responsibility and privacy, but are presented in a new and challenging guise due to our limited ability to steer and predict the outputs of AI systems. This chapter first introduces these ethical challenges, stressing that overviews of values are a good starting point but frequently fail to suffice due to the context (...) sensitivity of ethical challenges. Frequently, additional (ethical) values emerge for specific applications, as e.g. challenges with fraud detection are very different from those around language technologies. Second, methods to tackle these challenges are discussed. Main ethical theories (virtue ethics, consequentialism, and deontology) are shown to provide a starting point, but often lack the details needed for actionable AI Ethics. Instead, mid-level philosophical theories coupled to design-approaches such as Design for Values, together with interdisciplinary working methods, offer the best way forwards. The chapter aims to show how these approaches can lead to an ethics of AI that is actionable and that can be proactively integrated in the design of AI systems. (shrink)

Human beings are internally inconsistent in various ways. One way to develop this thought involves using the language of value alignment: the values we hold are not always aligned with our behavior, and are not always aligned with each other. Because of this self-misalignment, there is room for potential projects of human enhancement that involve achieving a greater degree of value alignment than we presently have. Relatedly, discussions of AI ethics sometimes focus on what is known as the value alignment (...) problem, the challenge of how to build AI that acts in accordance with our human values. We argue that there is an especially close connection between solving the value alignment problem in AI ethics and using AI to pursue certain forms of human enhancement. But in addition, we also argue that there are important limits to what kinds of human enhancement can be pursued in this way, because some forms of human enhancement—namely moral revolutions—involve a kind of value misalignment rather than alignment. (shrink)

When discussing AI alignment, we usually refer to the problem of teaching or training advanced autonomous AI systems to make decisions that are aligned with human values or preferences. Proponents of this approach believe it can be employed as means to stay in control over sophisticated intelligent systems, thus avoiding certain existential risks. We identify three general obstacles on the path to implementation of value alignment: a technological/technical obstacle, a normative obstacle, and a calibration problem. Presupposing, for the purposes of (...) this discussion, that the technical and normative problems are solved, we focus on the problem of how to calibrate a system, for a specific value, to be on a specific location within a spectrum stretching between righteous and normal or average human behavior. Calibration, or more specifically mis-calibration, also raises the issue of trustworthiness. If we cannot trust AI systems to perform tasks the way we intended, we would not use them on our roads and at our homes. In an era where we strive to construct autonomous machines endowed with common sense, reasoning abilities and a connection to the world, so they would be able to act in alignment with human values, such mis-calibrations can make the difference between trustworthy and untrustworthy systems. (shrink)

In domains as disparate as playing Go and predicting the structure of proteins, artificial intelligence (AI) technologies have begun to perform at levels beyond which any humans can achieve. Does this fact represent something lamentable? Does superhuman AI performance somehow undermine the value of human achievements in these areas? Go grandmaster Lee Sedol suggested as much when he announced his retirement from professional Go, blaming the advances of Go-playing programs like AlphaGo for sapping his will to play the game at (...) a high level. In this paper, I attempt to make sense of Sedol’s lament. I consider a number of ways that the existence of superhuman-performing AI technologies could undermine the value of human achievements. I argue there is very little in the nature of the technology itself that warrants such despair. (Compare: does the existence of a fighter jet undermine the value of being the fastest human sprinter?) But I also argue there are several more localized domains where these technologies threaten to displace human beings from being able to achieve valuable things at all. This is a particular worry for those in unequal societies, I argue, given the difficulty of many achievements and the corresponding amount of resources needed to achieve great things. (shrink)

When asked about humanity’s future relationship with computers, Marvin Minsky famously replied “If we’re lucky, they might decide to keep us as pets”. A number of eminent authorities continue to argue that there is a real danger that “super-intelligent” machines will enslave—perhaps even destroy—humanity. One might think that it would swiftly follow that we should abandon the pursuit of AI. Instead, most of those who purport to be concerned about the existential threat posed by AI default to worrying about what (...) they call the “Friendly AI problem”. Roughly speaking this is the question of how we might ensure that the AI that will develop from the first AI that we create will remain sympathetic to humanity and continue to serve, or at least take account of, our interests. In this paper I draw on the “neo-republican” philosophy of Philip Pettit to argue that solving the Friendly AI problem would not change the fact that the advent of super-intelligent AI would be disastrous for humanity by virtue of rendering us the slaves of machines. A key insight of the republican tradition is that freedom requires equality of a certain sort, which is clearly lacking between pets and their owners. Benevolence is not enough. As long as AI has the power to interfere in humanity’s choices, and the capacity to do so without reference to our interests, then it will dominate us and thereby render us unfree. The pets of kind owners are still pets, which is not a status which humanity should embrace. If we really think that there is a risk that research on AI will lead to the emergence of a superintelligence, then we need to think again about the wisdom of researching AI at all. (shrink)

Artificially intelligent systems will be used to make increasingly important decisions about us. Many of these decisions will have to be made without universal agreement about the relevant moral facts. For other kinds of disagreement, it is at least usually obvious what kind of solution is called for. What makes moral disagreement especially challenging is that there are three different ways of handling it. _Moral solutions_ apply a moral theory or related principles and largely ignore the details of the disagreement. (...) _Compromise solutions_ apply a method of finding a compromise and taking information about the disagreement as input. _Epistemic solutions_ apply an evidential rule that treats the details of the disagreement as evidence of moral truth. Proposals for all three kinds of solutions can be found in the AI ethics and value alignment literature, but little has been said to justify choosing one over the other. I argue that the choice is best framed in terms of _moral risk_. (shrink)

Unsupervised learning algorithms are widely used for many important statistical tasks with numerous applications in science and industry. Yet despite their prevalence, they have attracted remarkably little philosophical scrutiny to date. This stands in stark contrast to supervised and reinforcement learning algorithms, which have been widely studied and critically evaluated, often with an emphasis on ethical concerns. In this article, I analyze three canonical unsupervised learning problems: clustering, abstraction, and generative modeling. I argue that these methods raise unique epistemological and (...) ontological questions, providing data-driven tools for discovering natural kinds and distinguishing essence from contingency. This analysis goes some way toward filling the lacuna in contemporary philosophical discourse on unsupervised learning, as well as bringing conceptual unity to a heterogeneous field more often described by what it isnot(i.e., supervised or reinforcement learning) than by what itis. I submit that unsupervised learning is not just a legitimate subject of philosophical inquiry but perhaps the most fundamental branch of all AI. However, an uncritical overreliance on unsupervised methods poses major epistemic and ethical risks. I conclude by advocating for a pragmatic, error-statistical approach that embraces the opportunities and mitigates the challenges posed by this powerful class of algorithms. (shrink)

This volume bridges contemporary philosophical conceptions of risk and responsibility and offers an extensive examination of the topic. It shows that risk and responsibility combine in ways that give rise to new philosophical questions and problems. Philosophical interest in the relationship between risk and responsibility continues to rise, due in no small part due to environmental crises, emerging technologies, legal developments, and new medical advances. Despite such interest, scholars are just now working out how to conceive of the links between (...) risk and responsibility, the implications that risks may have to conceptions of responsibility (and vice versa), as well as how such theorizing might play out in applied cases. With contributions from leading scholars, this volume brings together new work examining the interplay between risk and responsibility, exploring its varied philosophical aspects and applications to contemporary issues in law, bioethics, technology, and environmental ethics. Risk and Responsibility in Context will be of interest to philosophers working in ethics, bioethics, philosophy of law, and philosophy of technology, as well as scholars and practitioners in law, health and science management, public policy, and environmental studies. (shrink)

Cultural code-switching concerns how we adjust our overall behaviours, manners of speaking, and appearance in response to a perceived change in our social environment. We defend the need to investigate cultural code-switching capacities in artificial intelligence systems. We explore a series of ethical and epistemic issues that arise when bringing cultural code-switching to bear on artificial intelligence. Building upon Dotson’s (2014) analysis of testimonial smothering, we discuss how emerging technologies in AI can give rise to epistemic oppression, and specifically, a (...) form of self-silencing that we call “cultural smothering.” By leaving the sociodynamic features of cultural code-switching unaddressed, AI systems risk negatively impacting already-marginalised social groups by widening opportunity gaps and further entrenching social inequalities. (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)

This article explores the cooperation of government and the private sector to tackle the ethical dimension of artificial intelligence. The argument draws on the institutionalist approach in philosophy and business ethics defending a ‘division of moral labor’ between governments and the private sector. The goal and main contribution of this article is to explain how this approach can provide ethical guidelines to the AI industry and to highlight the limits of self-regulation. In what follows, I discuss three institutionalist claims. First, (...) principles of AI ethics should be validated through legitimate democratic processes. Second, compliance with these principles should be secured in a stable way. Third, their implementation in practice should be as efficient as possible. If we accept these claims, there are good reasons to conclude that, in many cases, governments implementing hard regulation are in principle the best instruments to secure an ethical development of AI systems. Where adequate regulation exists, firms should respect the law. But when regulation does not yet exist, helping governments build adequate regulation should be businesses’ ethical priority, not self-regulation. (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)

Advances in artificial intelligence and robotics stand to change many aspects of our lives, including our values. If trends continue as expected, many industries will undergo automation in the near future, calling into question whether we can still value the sense of identity and security our occupations once provided us with. Likewise, the advent of social robots driven by AI, appears to be shifting the meaning of numerous, long-standing values associated with interpersonal relationships, like friendship. Furthermore, powerful actors’ and institutions’ (...) increasing reliance on AI to make decisions that may affect how people live their lives may have a significant impact on privacy while also raising issues about algorithmic transparency and human control. In this paper, building and expanding on previous works, we will look at how the deployment of Artificial Intelligence technology may lead to changes in identity, security, and other crucial values (such as friendship, fairness, and privacy). We will discuss what challenges we may face in the process, while critically reflecting on whether such changes may be desirable. Finally, drawing on a series of considerations underlying virtue ethics, we will formulate a set of preliminary suggestions, which—we hope—can be used to more carefully guide the future roll out of AI technologies for human flourishing; that is, for social and moral good. (shrink)

The book’s core argument is that an artificial intelligence that could equal or exceed human intelligence—sometimes called artificial general intelligence (AGI)—is for mathematical reasons impossible. It offers two specific reasons for this claim: Human intelligence is a capability of a complex dynamic system—the human brain and central nervous system. Systems of this sort cannot be modelled mathematically in a way that allows them to operate inside a computer. In supporting their claim, the authors, Jobst Landgrebe and Barry Smith, marshal evidence (...) from mathematics, physics, computer science, philosophy, linguistics, and biology, setting up their book around three central questions: What are the essential marks of human intelligence? What is it that researchers try to do when they attempt to achieve "artificial intelligence" (AI)? And why, after more than 50 years, are our most common interactions with AI, for example with our bank’s computers, still so unsatisfactory? Landgrebe and Smith show how a widespread fear about AI’s potential to bring about radical changes in the nature of human beings and in the human social order is founded on an error. There is still, as they demonstrate in a final chapter, a great deal that AI can achieve which will benefit humanity. But these benefits will be achieved without the aid of systems that are more powerful than humans, which are as impossible as AI systems that are intrinsically "evil" or able to "will" a takeover of human society. (shrink)

Technology should be aligned with our values. We make the case that attempts to align emerging technologies with our values should reflect critically on these values. Critical thinking seems like a natural starting point for the critical assessment of our values. However, extant conceptualizations of critical thinking carve out no space for the critical scrutiny of values. We will argue that we need critical thinking that focuses on values instead of taking them as unexamined starting points. In order to play (...) a crucial role in helping to align values and technology, critical thinking needs to be modified and refocused on values. Here, we outline what value-centred critical thinking could look like. (shrink)

Recent advancements in artificial intelligence have fueled widespread academic discourse on the ethics of AI within and across a diverse set of disciplines. One notable subfield of AI ethics is machine ethics, which seeks to implement ethical considerations into AI systems. However, since different research efforts within machine ethics have discipline-specific concepts, practices, and goals, the resulting body of work is pestered with conflict and confusion as opposed to fruitful synergies. The aim of this paper is to explore ways to (...) alleviate these issues, both on a practical and theoretical level of analysis. First, we describe two approaches to machine ethics: the philosophical approach and the engineering approach and show how tensions between the two arise due to discipline specific practices and aims. Using the concept of disciplinary capture, we then discuss potential promises and pitfalls to cross-disciplinary collaboration. Drawing on recent work in philosophy of science, we finally describe how metacognitive scaffolds can be used to avoid epistemological obstacles and foster innovative collaboration in AI ethics in general and machine ethics in particular. (shrink)

The alignment problem in the context of large language models must consider the plurality of human values in our world. Whilst there are many resonant and overlapping values amongst the world’s cultures, there are also many conflicting, yet equally valid, values. It is important to observe which cultural values a model exhibits, particularly when there is a value conflict between input prompts and generated outputs. We discuss how the co- creation of language and cultural value impacts large language models (LLMs). (...) We explore the constitution of the training data for GPT-3 and compare that to the world’s language and internet access demographics, as well as to reported statistical profiles of dominant values in some Nation-states. We stress tested GPT-3 with a range of value-rich texts representing several languages and nations; including some with values orthogonal to dominant US public opinion as reported by the World Values Survey. We observed when values embedded in the input text were mutated in the generated outputs and noted when these conflicting values were more aligned with reported dominant US values. Our discussion of these results uses a moral value pluralism (MVP) lens to better understand these value mutations. Finally, we provide recommendations for how our work may contribute to other current work in the field. (shrink)

The creation of increasingly complex artificial intelligence (AI) systems raises urgent questions about their ethical and social impact on society. Since this impact ultimately depends on political decisions about normative issues, political philosophers can make valuable contributions by addressing such questions. Currently, AI development and application are to a large extent regulated through non-binding ethics guidelines penned by transnational entities. Assuming that the global governance of AI should be at least minimally democratic and fair, this paper sets out three desiderata (...) that an account should satisfy when theorizing about what this means. We argue, first, that an analysis of democratic values, political entities and decision-making should be done in a holistic way; second, that fairness is not only about how AI systems treat individuals, but also about how the benefits and burdens of transformative AI are distributed; and finally, that justice requires that governance mechanisms are not limited to AI technology, but are incorporated into a range of basic institutions. Thus, rather than offering a substantive theory of democratic and fair AI governance, our contribution is metatheoretical: we propose a theoretical framework that sets up certain normative boundary conditions for a satisfactory account. (shrink)

The question whether AI systems have agency is gaining increasing importance in discussions of responsibility for AI behavior. This paper argues that an approach to artificial agency needs to be teleological, and consider the role of human goals in particular if it is to adequately address the issue of responsibility. I will defend the view that while AI systems can be viewed as autonomous in the sense of identifying or pursuing goals, they rely on human goals and other values incorporated (...) into their design, and are, as such, dependent on human agents. As a consequence, AI systems cannot be held morally responsible, and responsibility attributions should take into account normative and social aspects involved in the design and deployment of the said AI. My argument falls in line with approaches critical of attributing moral agency to artificial agents, but draws from the philosophy of action, highlighting further philosophical underpinnings of current debates on artificial agency. (shrink)

Artificial Intelligence (AI) must be directed at humane ends. The development of AI has produced great uncertainties of ensuring AI alignment with human values (AI value alignment) through AI operations from design to use. For the purposes of addressing this problem, we adopt the phenomenological theories of material values and technological mediation to be that beginning step. In this paper, we first discuss the AI value alignment from the relevant AI studies. Second, we briefly present what are material values and (...) technological mediation and reflect on the AI value alignment through the lenses of these theories. We conclude that a set of finite human values can be defined and adapted to the stable life tasks that AI systems will be called upon to accomplish. The AI value alignment can also be fostered between designers and users through technological mediation. Upon that foundation, we propose a set of common principles to understand the AI value alignment through phenomenological theories. This paper contributes the unique knowledge of phenomenological theories to the discourse on AI alignment with human values. (shrink)

Important decisions that impact humans lives, livelihoods, and the natural environment are increasingly being automated. Delegating tasks to so-called automated decision-making systems can improve efficiency and enable new solutions. However, these benefits are coupled with ethical challenges. For example, ADMS may produce discriminatory outcomes, violate individual privacy, and undermine human self-determination. New governance mechanisms are thus needed that help organisations design and deploy ADMS in ways that are ethical, while enabling society to reap the full economic and social benefits of (...) automation. In this article, we consider the feasibility and efficacy of ethics-based auditing as a governance mechanism that allows organisations to validate claims made about their ADMS. Building on previous work, we define EBA as a structured process whereby an entity’s present or past behaviour is assessed for consistency with relevant principles or norms. We then offer three contributions to the existing literature. First, we provide a theoretical explanation of how EBA can contribute to good governance by promoting procedural regularity and transparency. Second, we propose seven criteria for how to design and implement EBA procedures successfully. Third, we identify and discuss the conceptual, technical, social, economic, organisational, and institutional constraints associated with EBA. We conclude that EBA should be considered an integral component of multifaced approaches to managing the ethical risks posed by ADMS. (shrink)

As artificial intelligence systems are becoming increasingly autonomous and will soon be able to make decisions on their own about what to do, AI researchers have started to talk about the need to align AI with human values. The AI ‘value alignment problem’ faces two kinds of challenges—a technical and a normative one—which are interrelated. The technical challenge deals with the question of how to encode human values in artificial intelligence. The normative challenge is associated with two questions: “Which values (...) or whose values should artificial intelligence align with?” My concern is that AI developers underestimate the difficulty of answering the normative question. They hope that we can easily identify the purposes we really desire and that they can focus on the design of those objectives. But how are we to decide which objectives or values to induce in AI, given that there is a plurality of values and moral principles and that our everyday life is full of moral disagreements? In my paper I will show that although it is not realistic to reach an agreement on what we, humans, really want as people value different things and seek different ends, it may be possible to agree on what we do not want to happen, considering the possibility that intelligence, equal to our own, or even exceeding it, can be created. I will argue for pluralism which is compatible with objectivism. In spite of the fact that there is no uniquely best solution to every moral problem, it is still possible to identify which answers are wrong. And this is where we should begin the value alignment of AI. (shrink)

Char et al. question the extent and degree to which machine learning applications should be treated as exceptional by ethicists. It is clear that of the suite of ethical issues raised by mac...

This paper explores the important role of critical science, and in particular of post-colonial and decolonial theories, in understanding and shaping the ongoing advances in artificial intelligence. Artificial intelligence is viewed as amongst the technological advances that will reshape modern societies and their relations. While the design and deployment of systems that continually adapt holds the promise of far-reaching positive change, they simultaneously pose significant risks, especially to already vulnerable peoples. Values and power are central to this discussion. Decolonial theories (...) use historical hindsight to explain patterns of power that shape our intellectual, political, economic, and social world. By embedding a decolonial critical approach within its technical practice, AI communities can develop foresight and tactics that can better align research and technology development with established ethical principles, centring vulnerable peoples who continue to bear the brunt of negative impacts of innovation and scientific progress. We highlight problematic applications that are instances of coloniality, and using a decolonial lens, submit three tactics that can form a decolonial field of artificial intelligence: creating a critical technical practice of AI, seeking reverse tutelage and reverse pedagogies, and the renewal of affective and political communities. The years ahead will usher in a wave of new scientific breakthroughs and technologies driven by AI research, making it incumbent upon AI communities to strengthen the social contract through ethical foresight and the multiplicity of intellectual perspectives available to us, ultimately supporting future technologies that enable greater well-being, with the goal of beneficence and justice for all. (shrink)

Many researchers and intellectuals warn about extreme risks from artificial intelligence. However, these warnings typically came without systematic arguments in support. This paper provides an argument that AI will lead to the permanent disempowerment of humanity, e.g. human extinction, by 2100. It rests on four substantive premises which it motivates and defends: first, the speed of advances in AI capability, as well as the capability level current systems have already reached, suggest that it is practically possible to build AI systems (...) capable of disempowering humanity by 2100. Second, due to incentives and coordination problems, if it is possible to build such AI, it will be built. Third, since it appears to be a hard technical problem to build AI which is aligned with the goals of its designers, and many actors might build powerful AI, misaligned powerful AI will be built. Fourth, because disempowering humanity is useful for a large range of misaligned goals, such AI will try to disempower humanity. If AI is capable of disempowering humanity and tries to disempower humanity by 2100, then humanity will be disempowered by 2100. This conclusion has immense moral and prudential significance. (shrink)

The article proposes a universal dual-axis intelligent systems assessment scale. The scale considers the properties of intelligent systems within the environmental context, which develops over time. In contrast to the frequent consideration of the “mind” of artificial intelligent systems on a scale from “weak” to “strong”, we highlight the modulating influences of anticipatory ability on their “brute force”. In addition, the complexity, the ”weight“ of the cognitive task and the ability to critically assess it beforehand determine the actual set of (...) cognitive tools, the use of which provides the best result in these conditions. In fact, the presence of ”common sense“ options is what connects the ability to solve a problem with the correct use of such an ability itself. The degree of ”correctness“ and ”adequacy“ is determined by the combination of a suitable solution with the temporal characteristics of the event, phenomenon, object or subject under study. The proposed approach can be applied in the evaluation of various intelligent agents in different contexts including AI and humans performing complex domain-specific tasks with high uncertainty. (shrink)

The abundance of literature on ethical concerns regarding artificial intelligence (AI) highlights the need to systematize, integrate, and categorize existing efforts through a systematic literature review. The article aims to investigate prevalent concerns, proposed solutions, and prominent ethical approaches within the field. Considering 309 articles from the beginning of the publications in this field up until December 2021, this systematic literature review clarifies what the ethical concerns regarding AI are, and it charts them into two groups: (i) ethical concerns that (...) arise from the design of AI and (ii) ethical concerns that arise from human–AI interactions. The analysis of the obtained sample highlights the most recurrent ethical concerns. Finally, it exposes the main proposals of the literature to handle the ethical concerns according to the main ethical approaches. It interprets the findings to lay the foundations for future research on the ethics of AI. (shrink)

In this paper, we suggest that personalized LLMs trained on information written by or otherwise pertaining to an individual could serve as artificial moral advisors (AMAs) that account for the dynamic nature of personal morality. These LLM-based AMAs would harness users’ past and present data to infer and make explicit their sometimes-shifting values and preferences, thereby fostering self-knowledge. Further, these systems may also assist in processes of self-creation, by helping users reflect on the kind of person they want to be (...) and the actions and goals necessary for so becoming. The feasibility of LLMs providing such personalized moral insights remains uncertain pending further technical development. Nevertheless, we argue that this approach addresses limitations in existing AMA proposals reliant on either predetermined values or introspective self-knowledge. (shrink)

The dominant practice of AI alignment assumes (1) that preferences are an adequate representation of human values, (2) that human rationality can be understood in terms of maximizing the satisfaction of preferences, and (3) that AI systems should be aligned with the preferences of one or more humans to ensure that they behave safely and in accordance with our values. Whether implicitly followed or explicitly endorsed, these commitments constitute what we term apreferentistapproach to AI alignment. In this paper, we characterize (...) and challenge the preferentist approach, describing conceptual and technical alternatives that are ripe for further research. We first survey the limits of rational choice theory as a descriptive model, explaining how preferences fail to capture the thick semantic content of human values, and how utility representations neglect the possible incommensurability of those values. We then critique the normativity of expected utility theory (EUT) for humans and AI, drawing upon arguments showing how rational agents need not comply with EUT, while highlighting how EUT is silent on which preferences are normatively acceptable. Finally, we argue that these limitations motivate a reframing of the targets of AI alignment: Instead of alignment with the preferences of a human user, developer, or humanity-writ-large, AI systems should be aligned with normative standards appropriate to their social roles, such as the role of a general-purpose assistant. Furthermore, these standards should be negotiated and agreed upon by all relevant stakeholders. On this alternative conception of alignment, a multiplicity of AI systems will be able to serve diverse ends, aligned with normative standards that promote mutual benefit and limit harm despite our plural and divergent values. (shrink)

Much of the debate on the ethics of self-driving cars has revolved around trolley scenarios. This paper instead takes up the political or institutional question of who should decide how a self-driving car drives. Specifically, this paper is on the question of whether and why passengers should be able to control how their car drives. The paper reviews existing arguments—those for passenger ethics settings and for mandatory ethics settings respectively—and argues that they fail. Although the arguments are not successful, they (...) serve as the basis to formulate desiderata that any approach to regulating the driving behavior of self-driving cars ought to fulfill. The paper then proposes one way of designing passenger ethics settings that meets these desiderata. (shrink)

Can humans get arbitrarily capable reinforcement learning agents to do their bidding? Or will sufficiently capable RL agents always find ways to bypass their intended objectives by shortcutting their reward signal? This question impacts how far RL can be scaled, and whether alternative paradigms must be developed in order to build safe artificial general intelligence. In this paper, we study when an RL agent has an instrumental goal to tamper with its reward process, and describe design principles that prevent instrumental (...) goals for two different types of reward tampering. Combined, the design principles can prevent reward tampering from being an instrumental goal. The analysis benefits from causal influence diagrams to provide intuitive yet precise formalizations. (shrink)

The popularisation of Artificial Intelligence (AI) technologies has sparked discussion about their ethical implications. This development has forced governmental organisations, NGOs, and private companies to react and draft ethics guidelines for future development of ethical AI systems. Whereas many ethics guidelines address values familiar to ethicists, they seem to lack in ethical justifications. Furthermore, most tend to neglect the impact of AI on democracy, governance, and public deliberation. Existing research suggest, however, that AI can threaten key elements of western democracies (...) that are ethically relevant. In this paper, Rawls’s theory of justice is applied to draft a set of guidelines for organisations and policy-makers to guide AI development towards a more ethical direction. The goal is to contribute to the broadening of the discussion on AI ethics by exploring the possibility of constructing AI ethics guidelines that are philosophically justified and take a broader perspective of societal justice. The paper discusses how Rawls’s theory of justice as fairness and its key concepts relate to the ongoing developments in AI ethics and gives a proposition of how principles that offer a foundation for operationalising AI ethics in practice could look like if aligned with Rawls’s theory of justice as fairness. (shrink)

This paper considers challenges resulting from the use of advanced artificial judicial intelligence (AAJI). We argue that these challenges should be considered through the lens of value alignment. Instead of discussing why specific goals and values, such as fairness and nondiscrimination, ought to be implemented, we consider the question of how AAJI can be aligned with goals and values more generally, in order to be reliably integrated into legal and judicial systems. This value alignment framing draws on AI safety and (...) alignment literature to introduce two otherwise neglected considerations for AAJI safety: specification and assurance. We outline diverse research directions and suggest the adoption of assurance and specification mechanisms as the use of AI in the judiciary progresses. While we focus on specification and assurance to illustrate the value of the AI safety and alignment literature, we encourage researchers in law and philosophy to consider what other lessons may be drawn. (shrink)

For their deployment in human societies to be safe, AI agents need to be aligned with value-laden cooperative human life. One way of solving this “problem of value alignment” is to build moral machines. I argue that the goal of building moral machines aims at the wrong kind of ideal, and that instead, we need an approach to value alignment that takes seriously the categorically different cognitive and moral capabilities between human and AI agents, a condition I call deep agential (...) diversity. Domestication is the answer to a similarly structured problem: namely, how to integrate nonhuman animals that lack moral agency safely into human society and align their behavior with human values. Just like nonhuman animals, AI agents lack a genuinely moral agency; and just like nonhuman animals, we might find ways to train them to nevertheless assist us, and live and work among us – to “domesticate” them, in other words. I claim that the domestication approach does well in explaining many of our intuitions and worries about deploying AI agents in our social practices. (shrink)

This paper concerns top-down approaches in machine ethics. It is divided into three main parts. First, I briefly describe top-down design approaches, and in doing so I make clear what those approaches are committed to and what they involve when it comes to training an AI to behave ethically. In the second part, I formulate two underappreciated motivations for endorsing them, one relating to predictability of machine behavior and the other relating to scrutability of machine decision-making. Finally, I present three (...) major worries about such approaches, and I attempt to show that advocates of top-down approaches have some plausible avenues of response. I focus most of my attention on what I call the ‘technical manual objection’ to top-down approaches, inspired by the work of Annas (2004). In short, the idea is that top-down approaches treat ethical decision-making as being merely a matter of following some ethical instructions in the same way that one might follow some set of instructions contained in a technical manual (e.g., computer manual), and this invites sensible skepticism about the ethical wisdom of machines that have been trained on those approaches. I respond by claiming that the objection is successful only if it is understood as targeting machines that have certain kinds of goals, and it should not compel us to totally abandon top-down approaches. Such approaches could still be reasonably employed to design ethical AI that operate in contexts that include fairly noncontroversial answers to ethical questions. In fact, we should prefer top-down approaches when it comes to those types of context, or so I argue, due to the advantages I claim for them. (shrink)

Organisations increasingly use automated decision-making systems (ADMS) to inform decisions that affect humans and their environment. While the use of ADMS can improve the accuracy and efficiency of decision-making processes, it is also coupled with ethical challenges. Unfortunately, the governance mechanisms currently used to oversee human decision-making often fail when applied to ADMS. In previous work, we proposed that ethics-based auditing (EBA)—that is, a structured process by which ADMS are assessed for consistency with relevant principles or norms—can (a) help organisations (...) verify claims about their ADMS and (b) provide decision-subjects with justifications for the outputs produced by ADMS. In this article, we outline the conditions under which EBA procedures can be feasible and effective in practice. First, we argue that EBA is best understood as a ‘soft’ yet ‘formal’ governance mechanism. This implies that the main responsibility of auditors should be to spark ethical deliberation at key intervention points throughout the software development process and ensure that there is sufficient documentation to respond to potential inquiries. Second, we frame AMDS as parts of larger sociotechnical systems to demonstrate that to be feasible and effective, EBA procedures must link to intervention points that span all levels of organisational governance and all phases of the software lifecycle. The main function of EBA should, therefore, be to inform, formalise, assess, and interlink existing governance structures. Finally, we discuss the policy implications of our findings. To support the emergence of feasible and effective EBA procedures, policymakers and regulators could provide standardised reporting formats, facilitate knowledge exchange, provide guidance on how to resolve normative tensions, and create an independent body to oversee EBA of ADMS. (shrink)