Assesses whether LLMs used for summarising deliberative content, aggregating opinion, and predicting voter preferences genuinely advance democratic values or merely simulate democratic processes. Argues that optimism outpaces evidence, and that careful normative analysis is needed before deploying LLMs in democratic institutions.

Proposes Resource-Rational Contractualism (RRC) as an alignment framework: AI systems approximate the agreements rational parties would form using normatively-grounded, cognitively-inspired heuristics that trade effort for accuracy. An RRC-aligned agent operates efficiently while adapting dynamically to the changing human social world.

Identifies three shortcomings in existing LLM moral competence evaluations: over-reliance on prepackaged scenarios, focus on verdict prediction over reasoning, and inadequate testing of when models recognise they need more information. Introduces a five-dimensional assessment grounded in philosophical research on moral skill, covering perception, reasoning, self-knowledge, action, and reflection.

Proposes a richer framework for evaluating ethical competence in LLMs that goes beyond binary verdict accuracy to assess the quality of moral reasoning.

Autonomous AI cyber-weapons -- Military-AI Cyber Agents (MAICAs) -- create a credible pathway to catastrophic risk, given their technical feasibility and the geopolitical dynamics of cyberspace. Political, defensive-AI, and analogue-resilience measures are proposed to blunt the threat.

Develops a compound measure of how AI actors' power shapes the AI ecosystem, providing an empirical tool for tracking concentration and disparity.

Makes a moral case for deploying language model agents in recommendation systems, arguing they could better serve user interests than current algorithmic approaches.

Examines how AI agents may affect democratic institutions and proposes strategies for ensuring democratic resilience in the face of increasingly autonomous AI systems.

Argues for an anticipatory approach to AI ethics that addresses the societal impacts of emerging AI capabilities before they are widely deployed.

Argues that making AI agents safe and useful requires not just training-time behavioural modifications but external infrastructure -- including efficient agent-to-agent communication, attribution of actions to legal entities, and mechanisms for trust and accountability. Proposes an infrastructure-centric research agenda covering identification, data access, metering, and safety standards.

Warns that if dominant internet platforms control language model agents, the resulting 'platform agents' will deepen surveillance, tighten lock-in, and entrench incumbents. Advocates instead for user-controlled 'agent advocates,' requiring broad public access to compute and models, open interoperability standards, and market regulation to prevent platform foreclosure.

Assesses whether LLMs used for summarising deliberative content, aggregating opinion, and predicting voter preferences genuinely advance democratic values or merely simulate democratic processes. Argues that optimism outpaces evidence, and that careful normative analysis is needed before deploying LLMs in democratic institutions.

Drawing on Dewey, argues that computing technologies have transformed political association analogously to industrialisation, and develops a framework for governing the resulting algorithmic city.

Identifies five distinctive properties of open foundation models -- including greater customizability and poor monitoring -- that drive both their benefits and risks. Finds that current open models pose limited marginal risk of misuse beyond existing tools, but warns this calculus may shift as capabilities improve.

Analyses how algorithmic recommendation systems affect the development and exercise of moral skill, focusing on the role of attention in ethical perception.

Examines where justice demands intervene in predictive systems, analysing the appropriate locus of moral evaluation for algorithmic prediction.

Develops an account of democratic duties of explanation grounded in political legitimacy and authority, with implications for AI-driven decision-making.

Focuses on what makes generative AI systems genuinely distinctive rather than rehashing familiar AI pathologies or leaping to existential risk: their scientific achievement and the most consequential societal changes they will bring over the next decade. Explores the normative questions raised by language model agents that form the core of autonomous AI systems acting in the world.

Examines how ML and computational technologies create new forms of power over individuals, as corporations and diminished states exercise authority through automated systems.

Argues that algorithmic intermediaries govern the digital public sphere through architectures, amplification, and moderation, distributing attention in ways that raise questions of communicative justice.

Examines where justice demands intervene in predictive systems, analysing the appropriate locus of moral evaluation for algorithmic prediction.

Rapid, widespread adoption of the latest large language models has sparked both excitement and concern about advanced artificial intelligence (AI). In response, many are looking to the field of AI safety for answers. Major AI companies are purportedly investing heavily in this young research program, even as they cut "trust and safety" teams addressing harms from current systems. Governments are taking notice too. The United Kingdom just invested £100 million in a new "Foundation Model Taskforce" and plans an AI safety summit this year. And yet, as research priorities are being set, it is already clear that the prevailing technical agenda for AI safety is inadequate to address critical questions. Only a sociotechnical approach can truly limit current and potential dangers of advanced AI.

This paper examines three approaches to the relationship between our moral reasons to bear costs for others’ sake before and beyond the call of duty. Symmetry holds that you are required to optimise your beneficial sacrifices even when they are genuinely supererogatory. If you are required to bear a cost C for the sake of a benefit B, when they are the only costs and benefits at stake, you are also conditionally required to bear an additional cost C, for the sake of an additional benefit B, when enough other costs and benefits are at stake that both of your alternatives are presumptively supererogatory. Disconnection rejects the requirement to optimise when your options are presumptively supererogatory and maintains that you have an entirely free hand to choose as you will among them. Asymmetry holds that when acting beyond the call of duty you are entitled to a measure of additional freedom compared to when you are not taking on supererogatory costs—you can prioritise your own well-being and reasons to a greater degree—but places constraints on the options that you may permissibly choose. We defend a version of Asymmetry and explore its implications for recent debates on charitable giving.

Examines three central objections to Automated Influence -- privacy, exploitation, and manipulation -- showing that structural versions of each objection carry more weight than their interactional counterparts. By moving from the interactional focus of 'AI Ethics' to political philosophy, reveals that Automated Influence's core problem is the crisis of legitimacy it precipitates.

AI and related computational systems are being used by some to exercise power over others. This is especially clear in our online lives, which are increasingly structured and governed by computational systems using some of the most advanced techniques in AI. But it is also apparent in our offline lives, as computational systems using AI are used by powerful actors, including states, local government, and employers. Proponents of various principles of “AI Ethics” sometimes imply that the sole normative function of those principles is to ensure that AI is used to achieve socially acceptable goals. Drawing attention to the ways in which AI systems are used to exercise power demonstrates the inadequacy of this normative analysis. When new and intensified power relations develop, we must attend not only to what power is used for, but also to how and by whom it is used.

Many philosophers believe in lesser-evil justifications for doing harm: if the only way to stop a trolley from killing five is to divert it away onto one, then we may divert. But recently, Helen Frowe has argued that we do not only have the option to pursue the lesser evil: in most cases, we are so obligated. After critically assessing Frowe’s argument, I develop three mutually compatible accounts of lesser-evil options, which permit, but do not obligate us to minimize harm. These are the Parity Account, the Prerogative Account, and the Permissible Moral Mistakes Account. Considerations of parity and prerogatives have arisen in this debate before, but in inchoate form. The Permissible Moral Mistakes Account introduces something new.

Deontologists have avoided decision-making under uncertainty because standard decision theory appears superficially consequentialist, but its central tenets pose no genuine obstacle. Maps the path toward a comprehensive deontological decision theory that respects the priority of the right over the good.

If we had perfect information, then we could say, for any given objectively permissible act, what makes it objectively permissible. But when we have imperfect information, when we must decide under risk and uncertainty, what then makes an act <italic>subjectively</italic> permissible or impermissible? There are two salient possibilities. The first is the “verdicts” approach. It grounds judgments of subjective permissibility in probabilistically discounted judgments of objective permissibility. The principle “minimize expected objective wrongness” takes this approach. The second is the “reasons” approach. It grounds subjective permissibility in probabilistically discounted objective reasons. “Maximize expected utility” is one example. Chapter 10 considers whether the verdicts approach or the reasons approach to grounding judgments of subjective permissibility is better suited for <italic>deontological</italic> decision-making with imperfect information. Perhaps surprisingly, the reasons approach comes out on top.

:I argue that agent-centered options to favor and sacrifice one’s own interests are grounded in a particular aspect of self-ownership. Because you own your interests, you are entitled to a say over how they are used. That is, whether those interests count for or against some action is, at least in part, to be determined by your choice. This is not the only plausible argument for agent-centered options. But it has some virtues that other arguments lack.

Many of us think we have agent‐centred options to act suboptimally. Some of these involve favouring our own interests. Others involve sacrificing them. In this paper, I explore three different ways to accommodate agent‐centred options in a criterion of objective permissibility. I argue against satisficing and rational pluralism and in favour of a principle built around sensitivity to personal cost.

If we were required to sacrifice our own interests whenever doing so was best overall, or prohibited from doing so unless it was optimal, then we would be mere sites for the realization of value. Our interests, not ourselves, would wholly determine what we ought to do. We are not mere sites for the realization of value – instead we, ourselves, matter unconditionally. So we have options to act suboptimally. These options have limits, grounded in the very same considerations. Though not merely such sites, you and I are also sites for the realization of value, and our interests (and ourselves) must therefore sometimes determine what others ought to do, in particular requiring them to bear reasonable costs for our sake. Likewise, just as my moral status grounds a requirement that others show me appropriate respect, so must I do to myself.

Consider the following claim: given the choice between saving a life and preventing any number of people from temporarily experiencing a mild headache, you should always save the life. Many moral theorists accept this claim. In doing so, they commit themselves to some form of ‘moral absolutism’: the view that there are some moral considerations (like being able to save a life) that cannot be outweighed by any number of lesser moral considerations (like being able to avert a mild headache). In contexts of certainty, it is clear what moral absolutism requires of you. However, what does it require of you when deciding under risk? What ought you to do when there is a chance that, say, you will not succeed in saving the life? In recent years, various critics have argued that moral absolutism cannot satisfactorily deal with risk and should, therefore, be abandoned. In this paper, we show that moral absolutism can answer its critics by drawing on—of all things—orthodox expected utility theory.

Deontological ethical theory and unqualified absolutism were once near synonyms: 'do justice though the heavens fall'. Few deontologists today are so hard-line. But many still believe that some trade-offs that would yield unambiguously better outcomes are nonetheless wrong. Here is one such scenario-call it Life for Headaches: We must choose either to avert a minor, temporary headache for each member of a multitude or to save one person's life. 2 No matter how numerous the multitude, we ought always to save the one.

The problem of moral sunk costs pervades decision-making with respect to war. In the terms of just war theory, it may seem that incurring a large moral cost results in permissiveness: if a just goal may be reached at a small cost beyond that which was deemed proportionate at the outset of war, how can it be reasonable to require cessation? On this view, moral costs already expended could have major implications for the ethics of conflict termination. Discussion of sunk costs in moral theorizing about war has settled into four camps: Quota, Prospect, Addition, and Discount. In this paper, I offer a mathematical model that articulates each of these views. The purpose of the mathematisation is threefold. First, to unify the sunk costs problem. Second, to show that these views differ in the nature of their justifications: some are justified qualitatively and others quantitatively. Third, to clarify the differential force of qualitative and quantitative critiques of these four views.

Surveys two approaches to deontological decision-making under uncertainty about killing: using double-effect reasoning directly as a guide to subjective permissibility, and fitting deontological ethics into decision-theoretic apparatus. Shows the first approach faces insurmountable obstacles while the second holds much more promise than deontologists or their critics have assumed.

It is important to reflect on the way we evaluate the laws and customs of armed conflict and the responsibilities we take on when we criticize and propose possible changes to them. These laws are not robust, and there is a danger that criticism may undermine their force while not providing effective alternatives. Moreover, in the area of armed conflict, it is easy to underestimate the pressures that a satisfactory set of norms has to respond to and easy to exaggerate the “merely” conventional character of such norms. Laws of war must be administrable in circumstances of fear, confusion, and violence and must include elements of technicality difficult to understand in philosophical terms. One of the most influential of recent laws of war revisionists, Jeff McMahan, acknowledges that his deep moral critique of existing norms of armed conflict does not necessarily yield a set of prescriptions for legal reform. This chapter extends McMahan’s and counsels the utmost caution in these critiques and re-examinations.

This chapter introduces the two main ways to think about the ethics of war. The first is to start by thinking about war. The second is to think about the ethics of killing outside of war, then apply those principles to the case of war. In contemporary just war theory, the first approach has most commonly been associated with those who broadly aim to vindicate international law, such as Michael Walzer and his contemporary defenders. The second approach is more frequently linked to the work of Jeff McMahan, and Walzer’s other revisionist critics. I show that this conflation is mere accident. Indeed, perhaps the richest terrain to be ploughed is in the combinations that have been relatively neglected—vindications of international law that start from cases based outside of war; critiques of international law based on the distinctive nature of war.

Deontologists have long been upbraided for lacking an account of justified decision-making under risk and uncertainty. One response is to develop a deontological decision theory—a set of necessary and sufficient conditions for an act’s being permissible given an agent’s imperfect information. In this article, I show that deontologists can make more use of regular decision theory than some might have thought, but that we must adapt decision theory to accommodate agent-centered options—permissions to favor or sacrifice our own interests, when doing so is overall morally worse. Accommodating options requires more than just amending the decision-theoretic ‘value function’. We must change the decision rule as well.

Interest in just war theory has boomed in recent years, as a revisionist school of thought has challenged the orthodoxy of international law, most famously defended by Michael Walzer [1977]. These revisionist critics have targeted the two central principles governing the conduct of war (jus in bello): combatant equality and noncombatant immunity. The first states that combatants face the same permissions and constraints whether their cause is just or unjust. The second protects noncombatants from intentional attack. In response to these critics, some philosophers have defended aspects of the old orthodoxy on novel grounds. Revisionists counter. As things stand, the prospects for progress are remote. In this paper, we offer a way forward. We argue that exclusive focus on first-order moral principles, such as combatant equality and noncombatant immunity, has led revisionist and orthodox just war theorists to engage in “proxy battles.” Their first-order moral disagreements are at least partly traceable to second-order disagreements about the nature and purpose of political theory. These deeper disputes have been central to the broader discipline of political theory for several years; we hope that bringing them to bear on the ethics of war will help us move beyond the present impasse.

Situates the revisionist critique of Walzer's just war principles -- noncombatant immunity, proportionality, and combatant equality -- and illustrates the disturbing implications that follow from revisionist premises. Concludes that while revisionist arguments are powerful, alternative foundations for key tenets of international humanitarian law remain available.

Maps the contemporary divide between revisionist and traditionalist just war theory, from Walzer's foundational defence of international humanitarian law through McMahan's revisionist critique to the subsequent traditionalist revival. Shows that although revisionism identifies genuine moral problems, traditionalists can accommodate its strongest insights without abandoning the laws of armed conflict's moral foundations.

How should deontologists approach decision-making under uncertainty, for an iterated decision problem? In this article I explore the shortcomings of a simple expected value approach, using a novel example to raise questions about attitudes to risk, the moral significance of tiny probabilities, the independent moral reasons against imposing risks, the morality of sunk costs, and the role of agent-relativity in iterated decision problems.

Rejects the argument that unjust combatants are liable to be killed solely through complicity in their side's wrongful war, and that noncombatants escape targeting because they lack complicity. Argues instead that just combatants have reason to direct force against unjust combatants out of respect for the reasonable self-determining decisions of other political communities.

People often think that their special relationships with family, friends, comrades and compatriots, can ground moral reasons. Among these reasons, they understand some to be duties – pro tanto requirements that have genuine weight when they conflict with other considerations. In this paper I ask: what is the underlying moral structure of associative duties? I first consider and reject the orthodox Teleological Welfarist account, which first observes that special relationships are fundamental for human well-being, then claims that we cannot have these relationships, if we do not recognise associative duties, before concluding that we should therefore recognise associative duties. I then introduce a nonteleological alternative, grounded in the Appropriate Response approach to ethical theory.

Why does it matter that those who fight wars be authorized by the communities on whose behalf they claim to fight? I argue that lacking authorization generates a moral cost, which counts against a war's proportionality, and that having authorization allows the transfer of reasons from the members of the community to those who fight, which makes the war more likely to be proportionate. If democratic states are better able than non-democratic states and sub-state groups to gain their community's authorization, this means that some wars will be proportionate if fought by democracies, disproportionate if not.

Grounds part of the justification for killing in war in the associative duties we owe to loved ones -- to protect them from the severe harms war threatens. Shows how these duties can override the rights to life of those who must be killed, and how soldiers fighting on behalf of their community can act on reasons that apply to its members.

Killing civilians is worse than killing soldiers. Although this principle is widely affirmed, recent military practice and contemporary just war theory have undermined it. This article argues that killing an innocent person is worse the likelier it was, when you acted, that he would be innocent: riskier killings are worse than less risky killings. In war, killing innocent civilians is almost always riskier than killing innocent soldiers. So killing innocent civilians is worse than killing innocent soldiers. Since almost all civilians are innocent in war, and since killing innocent civilians is worse than killing liable soldiers, killing civilians is worse than killing soldiers.

The principle of noncombatant immunity prohibits warring parties from intentionally targeting noncombatants. I explicate the moral version of this view and its criticisms by reductive individualists; they argue that certain civilians on the unjust side are morally liable to be lethally targeted to forestall substantial contributions to that war. I then argue that reductivists are mistaken in thinking that causally contributing to an unjust war is a necessary condition for moral liability. Certain noncontributing civilians—notably, war-profiteers—can be morally liable to be lethally targeted. Thus, the principle of noncombatant immunity is mistaken as a moral (though not necessarily as a legal) doctrine, not just because some civilians contribute substantially, but because some unjustly enriched civilians culpably fail to discharge their restitutionary duties to those whose victimization made the unjust enrichment possible. Consequently, the moral criterion for lethal liability in war is even broader than reductive individualists have argued.

This chapter explains why one familiar and otherwise plausible approach to the justification of killing in war cannot adequately ground commonsense views of permissible national defence. Reductionists believe that justified warfare reduces to an aggregation of acts that are justified under ordinary principles of interpersonal morality. The standard form of reductionism focuses on the principles governing killing in ordinary life, specifically those that justify intentional killing in self- and other-defence, and unintended but foreseen (for short, collateral) killing as a lesser evil. Justified warfare, on this view, is no more than the coextension of multiple acts justified under these two principles. Reductionism is the default philosophical approach to thinking through the ethics of killing in war. It makes perfect sense to ask what principles govern permissible killing in general, before applying them to the particular context of war.

Grounds part of the justification for killing in war in the associative duties we owe to loved ones -- to protect them from the severe harms war threatens. Shows how these duties can override the rights to life of those who must be killed, and how soldiers fighting on behalf of their community can act on reasons that apply to its members.

War unleashes deadly violence and destruction and inverts the normal codes and conventions of society: killing, maiming, and injuring become justifiable; damage to land, resources, and buildings a normal occurrence; and the seizure of people and property a regularity. Following a century that has accustomed us to think about war in terms of totalitarian effects and total consequences, of genocidal campaigns, of the obliteration of cities and the threat of mutually assured destruction with atomic weapons, it may seem an arcane proclamation that war's reach and vehemence ought to be restrained. Yet that is what the just war conventions seek to explore and to uphold as guides to human action. While the pacifist rejects war absolutely and thereby contends that all acts within war are immoral and inexcusable and the militarist declares that in war (as in love) all is fair (Holmes, 1989), the just war theorist takes a middle position, acknowledging that should war break out, a range of considerations concerning its justification and the procedures that soldiers follow ought to be maintained as guides to ensure, in effect, that some sanctuaries from war's evils are acknowledged (Clark, 1988; Norman, 1995; Walzer, 1977).

Mainstream moral beliefs about war seem to be inconsistent with mainstream moral beliefs about self-defense such as the imminence requirement, the requirement to retreat, and restrictions on responses to conditional threats. The chapter argues that these apparent inconsistencies are actually the result of the necessity principle applied to environments with different nonmoral social facts. War takes place in the anarchy of international relations, where a lack of effective cosmopolitan security institutions makes it necessary to confront nonimminent threats, stand one’s ground, and respond to conditional threats in order to defeat and deter aggression. Self-defense, on the other hand, usually occurs in societies with effective police. In such an environment, it is unnecessary for private citizens to take similar actions to defeat and deter aggression. This is a good reason to believe that mainstream moral beliefs about war are closer to the true morality of war than recent critics claim.

The successful transition from armed conflict to peace is one of the greatest challenges of contemporary warfare. The laws and principles governing transitions from conflict to peace (jus post bellum) have only recently gained attention in legal scholarship. This volume investigates questions concerning the core of jus post bellum: the law (“jus”), the temporal aspect (“post”), and different types of armed conflict (“bellum”). It is the first volume to clarify the different legal meanings and components of the concept, including its implications in contemporary politics and practice. It explores the nature of jus post bellum as a concept, including its foundations, criticisms, and relationship to related concepts (e.g. Transitional Justice, Responsibility to Protect). It rethinks the nexus of the concept to jus ad bellum and jus in bello and its relevance in internal armed conflicts and peacebuilding. It examines problems in relation to the ending of conflict, including indicators for the end of conflict, exit strategies, and institutional responses. It also identifies contours of a “jus,” drawing on disparate bodies and sources of international law such as peace agreements, treaty law, self-determination, norms governing peace operations, and the status of foreign armed forces, environmental law, human rights, and amnesty law. Taking into account perspectives from multiple disciplines, the book will be relevant to scholars, practitioners, and students across many fields, such as peace and conflict studies, international relations, philosophy, political science, and international law.

The just war theory provides an account of rights and duties in war. At its heart lies a distinction between two sets of rules. The jus ad bellum provides a set of conditions under which going to war is deemed just. The jus in bello provides a set of conditions for just conduct in war. This chapter discusses the three challenges to just war theory and the responses to the three challenges.

This paper asks whether we can defend associative duties to our compatriots that are grounded solely in the relationship of liberal co‐citizenship. The sort of duties that are especially salient to this relationship are duties of justice, duties to protect and improve the institutions that constitute that relationship, and a duty to favour the interests of compatriots over those of foreigners. Critics have argued that the liberal conception of citizenship is too insubstantial to sustain these duties — indeed, that it gives us little reason to treat compatriots any differently from how we treat foreigners, with all the practical consequences that this would entail. I suggest that on a specific conception of liberal citizenship we can, in fact, defend associative duties, but that these extend only to the duty to protect and improve the institutions that constitute that relationship. Duties of justice and favouritism, I maintain, cannot be particularised to one's compatriots.

On one popular conception of how to do political theory, we should start with our considered judgments, try to work them together into a coherent theory, and then test our judgments against the theory, and the theory against the judgments, to see if either needs modification. 1 Philosophical discussion of the ethics of war has taken exactly this form: there are certain considered judgments, best enunciated by Michael Walzer, to which many hold. 2 Only combatants may be intentionally targeted in war; unintended harms to noncombatants must be minimized; wars of national defense and humanitarian intervention can be justified. Then there is the theory. Walzer's own loose attempt to synthesize these judgments has been largely discredited. 3 In recent years, philosophers from a more austere ethical tradition have argued that these theoretical failings demand reevaluation of the considered

In this chapter the main contemporary theories of justifiable killing in self-defense are criticized and an alternative account elaborated: the fault-based internalist suspendable-rights theory (FIST). FIST is a partialist account. The rights not to be killed are such that when one member of the set of rights is suspended, the other rights remain in force. Thus, if A’s right not to be killed by B is suspended, then B no longer has an obligation not to kill A. However, A still has a right not to be killed by C, and thus C’s obligation not to kill A remains in force. In addition, on FIST, a culpable attacker suspends his or her right not to be killed by a defender even in cases in which it is not necessary (necessity condition) for the defender to kill the attacker to save his or her own life.

Associative duties are the special obligations people are commonly thought to have toward certain associates, such as siblings, friends, colleagues, teammates, compatriots, and co‐nationals. While some doubt that associative duties are genuine, many take their existence to be necessary for explaining obligations which, upon reflection, appear difficult to ground in any other source.