There is a long history of the science of intelligent machines and its potential to provide scientific insights have been debated since the dawn of AI. In particular, there is renewed interest in the role of AI in research and research policy as an enabler of new methods, processes, management and evaluation which is still relatively under-explored. This empirical paper explores interviews with leading scholars on the potential impact of AI on research practice and culture through deductive, thematic analysis to (...) show the issues affecting academics and universities today. Our interviewees identify positive and negative consequences for research and researchers with respect to _collective_ and _individual use_. AI is perceived as helpful with respect to information gathering and other narrow tasks, and in support of impact and interdisciplinarity. However, using AI as a way of ‘speeding up—to keep up’ with bureaucratic and metricised processes, may proliferate negative aspects of academic culture in that the expansion of AI in research should assist and not replace human creativity. Research into the future role of AI in the research process needs to go further to address these challenges, and ask fundamental questions about how AI might assist in providing new tools able to question the values and principles driving institutions and research processes. We argue that to do this an explicit movement of meta-research on the role of AI in research should consider the effects for research and researcher creativity. Anticipatory approaches and engagement of diverse and critical voices at policy level and across disciplines should also be considered. (shrink)

Stories are an important indicator of our vision of the future. In the case of artificial intelligence, dominant stories are polarized between notions of threat and myopic solutionism. The central storytellers—big tech, popular media, and authors of science fiction—represent particular demographics and motivations. Many stories, and storytellers, are missing. This paper details the accounts of missing AI narratives by leading scholars from a range of disciplines interested in AI Futures. Participants focused on the gaps between dominant narratives and the untold (...) stories of the capabilities, issues, and everyday realities of the technology. One participant proposed a “story crisis” in which these narratives compete to shape the public discourse on AI. Our findings indicate that dominant narratives distract and mislead public understandings and conceptions of AI. This suggests a need to pay closer attention to missing AI narratives. It is not simply about telling new stories, it is about listening to existing stories and asking what is wanted from AI. We call for realistic, nuanced, and inclusive stories, working with and for diverse voices, which consider story-teller; genre, and communicative purpose. Such stories can then inspire the next generation of thinkers, technologists, and storytellers. (shrink)

Nash is famous for many inventions, but it is less known that he, simultaneously with Marschak, also was the first to axiomatize expected utility for risk. In particular, these authors were the first to state the independence condition, a condition that should have been but was not stated by von Neumann and Morgenstern. Marschak’s paper resulted from interactions with several people at the Cowles Commission. We document unique letters and personal communications with Nash, Samuelson, Arrow, Dalkey, and others, making plausible (...) that Nash made his discovery independently from the others. (shrink)

Think of a number, any number, or properties like fragility and humanity. These and other abstract entities are radically different from concrete entities like electrons and elbows. While concrete entities are located in space and time, have causes and effects, and are known through empirical means, abstract entities like meanings and possibilities are remarkably different. They seem to be immutable and imperceptible and to exist "outside" of space and time. This book provides a comprehensive critical assessment of the problems raised (...) by abstract entities and the debates about existence, truth, and knowledge that surround them. It sets out the key issues that inform the metaphysical disagreement between platonists who accept abstract entities and nominalists who deny abstract entities exist. Beginning with the essentials of the platonist–nominalist debate, it explores the key arguments and issues informing the contemporary debate over abstract reality: arguments for platonism and their connections to semantics, science, and metaphysical explanation the abstract–concrete distinction and views about the nature of abstract reality epistemological puzzles surrounding our knowledge of mathematical entities and other abstract entities. arguments for nominalism premised upon concerns about paradox, parsimony, infinite regresses, underdetermination, and causal isolation nominalist options that seek to dispense with abstract entities. Including chapter summaries, annotated further reading, and a glossary, _Entities_ is essential reading for anyone seeking a clear and authoritative introduction to the problems raised by abstract entities. (shrink)

Within contemporary science, it is common practice to compare data points to the average, i.e., to the statistical mean. Because this practice is so familiar, it might at first appear not to be the sort of thing that requires explanation. But recent research in cognitive science and in the history of science gives us reason to adopt the opposite perspective. Cognitive science research on the ways people ordinarily make sense of the world suggests that, instead of using a purely statistical (...) notion of the average, people tend to use a value-laden notion of the normal. Similarly, historical research on the scientific practices used in other time periods indicates that, prior to the nineteenth century, scientists tended to make use of a notion of the normal that was overtly value-laden. These findings give us reason to rethink certain familiar facts about scientific practice. In particular, they suggest that the fact that scientists so often make use of the statistical average should be seen as a highly surprising fact, the sort of thing that calls out for explanation. (shrink)

The idea of a single scientific method, shared across specialties and teachable to ten-year-olds, is just over a hundred years old. For centuries prior, science had meant a kind of knowledge, made from facts gathered through direct observation or deduced from first principles. But during the nineteenth century, science came to mean something else: a way of thinking. The Scientific Method tells the story of how this approach took hold in laboratories, the field, and eventually classrooms, where science was once (...) taught as a natural process. Henry M. Cowles reveals the intertwined histories of evolution and experiment, from Charles Darwin's theory of natural selection to John Dewey's vision for science education. Darwin portrayed nature as akin to a man of science, experimenting through evolution, while his followers turned his theory onto the mind itself. Psychologists reimagined the scientific method as a problem-solving adaptation, a basic feature of cognition that had helped humans prosper. This was how Dewey and other educators taught science at the turn of the twentieth century-but their organic account was not to last. Soon, the scientific method was reimagined as a means of controlling nature, not a product of it. By shedding its roots in evolutionary theory, the scientific method came to seem far less natural, but far more powerful. This book reveals the origin of a fundamental modern concept. Once seen as a natural adaptation, the method soon became a symbol of science's power over nature, a power that, until recently, has rarely been called into question. (shrink)

This article reports the findings of AI4People, an Atomium—EISMD initiative designed to lay the foundations for a “Good AI Society”. We introduce the core opportunities and risks of AI for society; present a synthesis of five ethical principles that should undergird its development and adoption; and offer 20 concrete recommendations—to assess, to develop, to incentivise, and to support good AI—which in some cases may be undertaken directly by national or supranational policy makers, while in others may be led by other (...) stakeholders. If adopted, these recommendations would serve as a firm foundation for the establishment of a Good AI Society. (shrink)

Artificial Intelligence (AI) is already having a major impact on society. As a result, many organizations have launched a wide range of initiatives to establish ethical principles for the adoption of socially beneficial AI. Unfortunately, the sheer volume of proposed principles threatens to overwhelm and confuse. How might this problem of ‘principle proliferation’ be solved? In this paper, we report the results of a fine-grained analysis of several of the highest-profile sets of ethical principles for AI. We assess whether these (...) principles converge upon a set of agreed-upon principles, or diverge, with significant disagreement over what constitutes ‘ethical AI.’ Our analysis finds a high degree of overlap among the sets of principles we analyze. We then identify an overarching framework consisting of five core principles for ethical AI. Four of them are core principles commonly used in bioethics: beneficence, non-maleficence, autonomy, and justice. On the basis of our comparative analysis, we argue that a new principle is needed in addition: explicability, understood as incorporating both the epistemological sense of intelligibility (as an answer to the question ‘how does it work?’) and in the ethical sense of accountability (as an answer to the question: ‘who is responsible for the way it works?’). In the ensuing discussion, we note the limitations and assess the implications of this ethical framework for future efforts to create laws, rules, technical standards, and best practices for ethical AI in a wide range of contexts. (shrink)

The idea of artificial intelligence for social good is gaining traction within information societies in general and the AI community in particular. It has the potential to tackle social problems through the development of AI-based solutions. Yet, to date, there is only limited understanding of what makes AI socially good in theory, what counts as AI4SG in practice, and how to reproduce its initial successes in terms of policies. This article addresses this gap by identifying seven ethical factors that are (...) essential for future AI4SG initiatives. The analysis is supported by 27 case examples of AI4SG projects. Some of these factors are almost entirely novel to AI, while the significance of other factors is heightened by the use of AI. From each of these factors, corresponding best practices are formulated which, subject to context and balance, may serve as preliminary guidelines to ensure that well-designed AI is more likely to serve the social good. (shrink)

As I write this, in November 1971, people are dying in East Bengal from lack of food, shelter, and medical caxc. The suffering and death that are occurring there now axe not inevitable, 1101; unavoidable in any fatalistic sense of the term. Constant poverty, a cyclone, and a civil war have turned at least nine million people into destitute refugees; nevertheless, it is not beyond Lhe capacity of the richer nations to give enough assistance to reduce any further suffering to (...) very small proportions. The decisions arid actions of human beings can prevent this kind of suffering. Unfortunately, human beings have not made the necessary decisions. At the individual level, people have, with very few exceptions, not responded to the situation in any signihczmt way. Generally speaking, people have not given large sums to relief funds; they have not written t0 their parliamentaxy representatives demanding increased government assistance; they have not demonstrated in the streets, held symbolic fasts, or done anything else directed toward providing thc refugees with the means to satisfy their essential needs. At thc government level, no govcrmmcnt has given the sort of massive aid that would enable the refugees to survive fm: more than a few days. Britain, for instance, has given rather more than most countries. It has, to date, given £14,750,000. For comparative purposes, B1~itain’s share of thc nonreccverabla development costs of the Anglo-French Concorde project is already in excess of £275,000,000, and on present estimates will reach £440,000,¤00. The implication is that the British government values a supersonic transport more than thirty times as.. (shrink)

In July 2017, China’s State Council released the country’s strategy for developing artificial intelligence, entitled ‘New Generation Artificial Intelligence Development Plan’. This strategy outlined China’s aims to become the world leader in AI by 2030, to monetise AI into a trillion-yuan industry, and to emerge as the driving force in defining ethical norms and standards for AI. Several reports have analysed specific aspects of China’s AI policies or have assessed the country’s technical capabilities. Instead, in this article, we focus on (...) the socio-political background and policy debates that are shaping China’s AI strategy. In particular, we analyse the main strategic areas in which China is investing in AI and the concurrent ethical debates that are delimiting its use. By focusing on the policy backdrop, we seek to provide a more comprehensive and critical understanding of China’s AI policy by bringing together debates and analyses of a wide array of policy documents. (shrink)

Research on the ethics of algorithms has grown substantially over the past decade. Alongside the exponential development and application of machine learning algorithms, new ethical problems and solutions relating to their ubiquitous use in society have been proposed. This article builds on a review of the ethics of algorithms published in 2016, 2016). The goals are to contribute to the debate on the identification and analysis of the ethical implications of algorithms, to provide an updated analysis of epistemic and normative (...) concerns, and to offer actionable guidance for the governance of the design, development and deployment of algorithms. (shrink)

Over the past few years, there has been a proliferation of artificial intelligence strategies, released by governments around the world, that seek to maximise the benefits of AI and minimise potential harms. This article provides a comparative analysis of the European Union and the United States’ AI strategies and considers the visions of a ‘Good AI Society’ that are forwarded in key policy documents and their opportunity costs, the extent to which the implementation of each vision is living up to (...) stated aims and the consequences that these differing visions of a ‘Good AI Society’ have for transatlantic cooperation. The article concludes by comparing the ethical desirability of each vision and identifies areas where the EU, and especially the US, need to improve in order to achieve ethical outcomes and deepen cooperation. (shrink)

Research on the ethics of algorithms has grown substantially over the past decade. Alongside the exponential development and application of machine learning algorithms, new ethical problems and solutions relating to their ubiquitous use in society have been proposed. This article builds on a review of the ethics of algorithms published in 2016, 2016). The goals are to contribute to the debate on the identification and analysis of the ethical implications of algorithms, to provide an updated analysis of epistemic and normative (...) concerns, and to offer actionable guidance for the governance of the design, development and deployment of algorithms. (shrink)

Technologies to rapidly alert people when they have been in contact with someone carrying the coronavirus SARS-CoV-2 are part of a strategy to bring the pandemic under control. Currently, at least 47 contact-tracing apps are available globally. They are already in use in Australia, South Korea and Singapore, for instance. And many other governments are testing or considering them. Here we set out 16 questions to assess whether — and to what extent — a contact-tracing app is ethically justifiable.

This paper argues that a set of questioning attitudes are among the foundations of human and animal minds. While both verbal questioning and states of curiosity are generally explained in terms of metacognitive desires for knowledge or true belief, I argue that each is better explained by a prelinguistic sui generis type of mental attitude of questioning. I review a range of considerations in support of such a proposal and improve on previous characterizations of the nature of these attitudes. I (...) then broaden their explanatory scope to include a number of forms of exploratory search. The paper has three main goals: to characterize the nature of the questioning attitudes, outlining their causal role and type of content; to argue that they are fundamental components of the mind, being widespread among animals and not constructed during ontogeny out of other attitudes; and to suggest that they explain a great deal more behavior than one might think. (shrink)

We propose a solution to the problem of logical omniscience in what we take to be its fundamental version: as concerning arbitrary agents and the knowledge attitude per se. Our logic of knowledge is a spin-off from a general theory of thick content, whereby the content of a sentence has two components: an intension, taking care of truth conditions; and a topic, taking care of subject matter. We present a list of plausible logical validities and invalidities for the logic of (...) knowledge per se for arbitrary agents, and isolate three explanatory factors for them: the topic-sensitivity of content; the fragmentation of knowledge states; the defeasibility of knowledge acquisition. We then present a novel dynamic epistemic logic that yields precisely the desired validities and invalidities, for which we provide expressivity and completeness results. We contrast this with related systems and address possible objections. (shrink)

What is ethics? Where do moral standards come from? Are they based on emotions, reason, or some innate sense of right and wrong? For many scientists, the key lies entirely in biology---especially in Darwinian theories of evolution and self-preservation. But if evolution is a struggle for survival, why are we still capable of altruism? In his classic study The Expanding Circle, Peter Singer argues that altruism began as a genetically based drive to protect one's kin and community members but (...) has developed into a consciously chosen ethic with an expanding circle of moral concern. Drawing on philosophy and evolutionary psychology, he demonstrates that human ethics cannot be explained by biology alone. Rather, it is our capacity for reasoning that makes moral progress possible. In a new afterword, Singer takes stock of his argument in light or recent research on the evolution of morality.--Publisher's description. (shrink)

Narrative thinking -- Narrative thinking about one's past -- Grief : a case study -- Narrative thinking about one's future -- Self-forgiveness : a case study -- The narrative sense of self -- Narrative, truth, life, and fiction.

Single-premise epistemic closure is the principle that: if one is in an evidential position to know that P where P entails Q, then one is in an evidential position to know that Q. In this paper, I defend the viability of opposition to closure. A key task for such an opponent is to precisely formulate a restricted closure principle that remains true to the motivations for abandoning unrestricted closure but does not endorse particularly egregious instances of closure violation. I focus (...) on two brands of epistemic theory that naturally incorporate closure restrictions. The first type holds that the truth value of a knowledge ascription is relative to a relevant question. The second holds that the truth value of a knowledge ascription is relative to a relevant topic. For each approach, I offer a formalization of a leading theory from the literature and use this formalization to evaluate the theory’s adequacy in terms of a precise set of desiderata. I conclude that neither theory succeeds in meeting these desiderata, casting doubt on the viability of the underlying approaches. Finally, I offer a novel variant of the topic-sensitive approach that fares better. (shrink)

This chapter provides an introduction to higher-order metaphysics as well as to the contributions to this volume. We discuss five topics, corresponding to the five parts of this volume, and summarize the contributions to each part. First, we motivate the usefulness of higher-order quantification in metaphysics using a number of examples, and discuss the question of how such quantifiers should be interpreted. We provide a brief introduction to the most common forms of higher-order logics used in metaphysics, and indicate a (...) number of questions which can be raised in such systems using logical vocabulary alone. Using a further example, we return to applications of higher-order logics in metaphysics. We also mention key developments in the history of higher-order logic as it pertains to metaphysics. Finally, we mention certain arguments which have been raised against the use of higher-order logic, and some ways of responding to them. (shrink)

A familiar phenomenon in the history of science is the transference of a concept from one field of investigation to another. At present the newer concepts in use in physics are being applied in the social sciences. The classic example of this transference phenomenon is the concept of evolution; originating in the social sciences, it was taken over into biology and thence to the whole range of current thought. This paper aims at tracing the use of one idea--the struggle for (...) existence--by Malthus in 1798 in an attempt to solve a problem in economics and politics, its elaboration by Darwin sixty years later into the theory of natural selection for the purpose of solving a problem in biology, and the immediate application of the combined doctrines by Walter Bagehot in an effort to solve the problem of social change. (shrink)

The World Health Organisation declared COVID-19 a global pandemic on 11th March 2020, recognising that the underlying SARS-CoV-2 has caused the greatest global crisis since World War II. In this chapter, we present a framework to evaluate whether and to what extent the use of digital systems that track and/or trace potentially infected individuals is not only legal but also ethical.

Introduction These essays have been written over a period of about ten years and have already been published separately in various places. ...

The theoretical virtue of parsimony values the minimizing of theoretical commitments, but theoretical commitments come in two kinds : ontological and ideological. While the ontological commitments of a theory are the entities it posits, a theory’s ideological commitments are the primitive concepts it employs. Here, I show how we can extend the distinction between quantitative and qualitative parsimony, commonly drawn regarding ontological commitments, to the domain of ideological commitments. I then argue that qualitative ideological parsimony is a theoretical virtue. My (...) defense proceeds by demonstrating the merits of qualitative ideological parsimony and by showing how the qualitative conception of ideological parsimony undermines two notable arguments from ideological parsimony: David Lewis’ defense of modal realism and Ted Sider’s defense of mereological nihilism. (shrink)

Initiatives relying on artificial intelligence (AI) to deliver socially beneficial outcomes—AI for social good (AI4SG)—are on the rise. However, existing attempts to understand and foster AI4SG initiatives have so far been limited by the lack of normative analyses and a shortage of empirical evidence. In this Perspective, we address these limitations by providing a definition of AI4SG and by advocating the use of the United Nations’ Sustainable Development Goals (SDGs) as a benchmark for tracing the scope and spread of AI4SG. (...) We introduce a database of AI4SG projects gathered using this benchmark, and discuss several key insights, including the extent to which different SDGs are being addressed. This analysis makes possible the identification of pressing problems that, if left unaddressed, risk hampering the effectiveness of AI4SG initiatives. (shrink)

First published in 1987. Routledge is an imprint of Taylor & Francis, an informa company.

Since its original publication in 1975, this groundbreaking work has awakened millions of people to the existence of "speciesism"—our systematic disregard of nonhuman animals—inspiring a worldwide movement to transform our attitudes to animals and eliminate the cruelty we inflict on them. In Animal Liberation, author Peter Singer exposes the chilling realities of today’s "factory farms" and product-testing procedures—destroying the spurious justifications behind them, and offering alternatives to what has become a profound environmental and social as well as moral issue. (...) An important and persuasive appeal to conscience, fairness, decency, and justice, it is essential reading for the supporter and the skeptic alike. (shrink)

I lay out the framework for my theory of sensory imagination in “Imagining as a guide to possibility.” Sensory imagining involves mental imagery , and crucially, in describing the content of imagining, I distinguish between qualitative content and assigned content. Qualitative content derives from the mental image itself; for visual imaginings, it is what is “pictured.” For example, visually imagine the Philadelphia Eagles defeating the Pittsburgh Steelers to win their first Super Bowl. You picture the greenness of the field and (...) the football’s brown oblong shape. Some of what you imagine isn’t explicitly pictured, however. That it is Sunday, that it is the Super Bowl: these facts are assigned. (shrink)

Political decisions have the potential to greatly impact our lives. Think of decisions in relation to abortion or climate change, for example. This makes political legitimacy an important normative concern. But what makes political decisions legitimate? Are they legitimate in virtue of having support from the citizens? Democratic conceptions of political legitimacy answer in the affirmative. Such conceptions righly highlight that legitimate political decision-making must be sensitive to disagreements among the citizens. But what if democratic decisions fail to track what (...) there is most reason to do? What if a democratically elected government fails to take measures necessary to protect its population from threats related to climate change? Peter argues that the legitimacy of political decisions doesn't just depend on respect for the citizens' will; and defends a novel hybrid conception of political legitimacy, called the Epistemic Accountability conception. According to this conception, political legitimacy also depends on how political decision-making responds to evidence for what there is most reason to do. Grounds for Political Legitmacy starts with an overview of the main ways in which philosophers have thought about political legitimacy, and identifies the epistemic accountability conception as an overlooked alternative. It then develops the epistemic accountability conception of political legitimacy and discusses its implications for legitimate political decision-making. Considering the norms that should govern political debate, it examines the role of experts in politics, and probes the responsibilities of democratically elected political leaders and as well as of citizens. (shrink)

According to the modal view, essence admits of reductive analysis in exclusively modal terms. Fine (1994) argues that modal view delivers an inadequate analysis of essence. This paper defends the modal view from Fine's challenge. This defense proceeds by examining the disagreement between Finean primitivists and Quinean eliminativists about essence. In order to model this disagreement, a distinction between essence and a separable concept, nature, is required. This distinction is then used to show that Fine's challenge is misdirected and therefore (...) unsuccessful. (shrink)

Discussion of atomistic and monistic theses about abstract reality.

That AI will have a major impact on society is no longer in question. Current debate turns instead on how far this impact will be positive or negative, for whom, in which ways, in which places, and on what timescale. In order to frame these questions in a more substantive way, in this prolegomena we introduce what we consider the four core opportunities for society offered by the use of AI, four associated risks which could emerge from its overuse or (...) misuse, and the opportunity costs associated with its under use. We then offer a high-level view of the emerging advantages for organisations of taking an ethical approach to developing and deploying AI. Finally, we introduce a set of five principles which should guide the development and deployment of AI technologies. The development of laws, policies and best practices for seizing the opportunities and minimizing the risks posed by AI technologies would benefit from building on ethical frameworks such as the one offered here. (shrink)

In this article we analyse the role that artificial intelligence (AI) could play, and is playing, to combat global climate change. We identify two crucial opportunities that AI offers in this domain: it can help improve and expand current understanding of climate change and it contribute to combating the climate crisis effectively. However, the development of AI also raises two sets of problems when considering climate change: the possible exacerbation of social and ethical challenges already associated with AI, and the (...) contribution to climate change of the greenhouse gases emitted by training data and computation-intensive AI systems. We assess the carbon footprint of AI research, and the factors that influence AI’s greenhouse gas (GHG) emissions in this domain. We find that the carbon footprint of AI research may be significant and highlight the need for more evidence concerning the trade-off between the GHG emissions generated by AI research and the energy and resource efficiency gains that AI can offer. In light of our analysis, we argue that leveraging the opportunities offered by AI for global climate change whilst limiting its risks is a gambit which requires responsive, evidence-based and effective governance to become a winning strategy. We conclude by identifying the European Union as being especially well-placed to play a leading role in this policy response and provide 13 recommendations that are designed to identify and harness the opportunities of AI for combating climate change, while reducing its impact on the environment. (shrink)

This volume explores the use of higher-order logics in metaphysics. Seventeen original essays trace the development of higher-order metaphysics, discuss different ways in which higher-order languages and logics may be used, and consider their application to various central topics of metaphysics.

In the Western world it is usually taken as given that we all want happiness, and our educational arrangements tacitly acknowledge this. Happiness, Hope, and Despair argues, however, that education has an important role to play in deepening our understanding of suffering and despair as well as happiness and joy. Education can be uncomfortable, unpredictable, and unsettling; it can lead to greater uncertainty and unhappiness. Drawing on the work of Søren Kierkegaard, Miguel de Unamuno, Fyodor Dostoevsky, Simone Weil, Paulo Freire, (...) and others, Peter Roberts shows why these features of educational life need not be feared; to the contrary, they can be seen as a source of hope and human fulfilment. After years of negotiating an education system dominated by the language of competition, performance, and economic advancement, students and teachers often long for something different; they seek not just measurable success but also opportunities to ask searching questions of themselves and the world they encounter. Happiness, Hope, and Despair makes an important contribution toward meeting this need. It fosters a rethinking of the nature, purpose, and value of education, and opens up possibilities for further scholarly and professional inquiry. (shrink)

The distinction between qualitative properties like mass and shape and non-qualitative properties like being Napoleon and being next to Obama is important, but remains largely unexamined. After discussing its theoretical significance and cataloguing various kinds of non-qualitative properties, I survey several views about the nature of this distinction and argue that all proposed reductive analyses of this distinction are unsatisfactory. I then defend primitivism, according to which the distinction resists reductive analysis.

In this article, we analyse the role that artificial intelligence (AI) could play, and is playing, to combat global climate change. We identify two crucial opportunities that AI offers in this domain: it can help improve and expand current understanding of climate change, and it can contribute to combatting the climate crisis effectively. However, the development of AI also raises two sets of problems when considering climate change: the possible exacerbation of social and ethical challenges already associated with AI, and (...) the contribution to climate change of the greenhouse gases emitted by training data and computation-intensive AI systems. We assess the carbon footprint of AI research, and the factors that influence AI’s greenhouse gas (GHG) emissions in this domain. We find that the carbon footprint of AI research may be significant and highlight the need for more evidence concerning the trade-off between the GHG emissions generated by AI research and the energy and resource efficiency gains that AI can offer. In light of our analysis, we argue that leveraging the opportunities offered by AI for global climate change whilst limiting its risks is a gambit which requires responsive, evidence-based, and effective governance to become a winning strategy. We conclude by identifying the European Union as being especially well-placed to play a leading role in this policy response and provide 13 recommendations that are designed to identify and harness the opportunities of AI for combatting climate change, while reducing its impact on the environment. (shrink)