Ryan Jenkins

In this chapter, we argue that deontological evaluations do not apply to the actions of robots. If robots are not phenomenally conscious, there is good reason to believe they lack the repertoire of mental capacities required for agency. Deontological requirements apply to a thing only if that thing is a moral agent. For this reason, robots should be consequentialists, even if consequentialism is false. We also argue that this does not necessarily make it permissible to create consequentialist robots, if we could instead create perfect deontological robots. But we are not sure about what true moral theory is, about appropriate trade-offs between rights and utilities, or about the ability of robots to obey moral commands reliably. So we argue that specific forms of uncertainty make it permissible, and sometimes obligatory, to create robots with moral views one thinks are false.

Is there an ideal war, a war morally better than which no war can be conceived? As long as the concept of an ideal war is coherent—as this chapter argues in this chapter—we should answer the opening question like this: An ideal war would be a war wherein civilian casualties were minimal or nonexistent and where acts of violence perfectly discriminated between combatants and noncombatants. Cyberwarfare has made possible this kind of ideal warfare for the first time by profoundly improving a state’s ability to direct its force discriminately and to ensure that force is proportional. Since cyberwarfare does not raise any moral concerns serious enough to countervail its clear benefits, we are obligated to prefer cybermeans where practical. These benefits of cyberwarfare undermine the moral stringency of the proportionality and probability of success criteria of _jus ad bellum_.

As police departments have come to rely on algorithmic patrol management systems to assign patrols, community groups and academics have raised concerns about demographic bias in the data used to train these systems. This paper introduces a technique for eliminating demographically correlated features from a place-based algorithmic patrol management system. We apply our technique to the real-world algorithmic patrol management system ResourceRouter, developed by SoundThinking. After applying the technique, we test the system's accuracy against historical crime incident data in four cities across several crime categories. Results suggest that demographically correlated features can be removed from the system without reducing accuracy.

Rule consequentialism is usually taken to recommend a single ideal code for all moral agents. Here I argue that, depending on their theoretical motivations, some rule consequentialists have good reasons to be relativists. Rule consequentialists who are moved by consequentialist considerations ought to support a scheme of multiple relativized moral codes because we could expect such a scheme to have better consequences in terms of impartial aggregate wellbeing than a single universal code. Rule consequentialists who find compelling the theory’s coherence with our considered moral intuitions should do the same because a scheme of multiple codes could better cohere with our intuitions about costless benefits, though these intuitions must be weighed against our allegiance to moral universalism.

This paper engages with the problem of toxic speech online and suggests remedies inspired by the value-sensitive design literature (VSD), suggesting that the designers of online platforms should explore methods of adding friction to online conversations. Second, this paper examines a historical case of designing a communications platform to offer methods to users to inculcate norms of acceptable behavior by introducing friction into synchronous conversations. This is the case of America Online (AOL) Instant Messenger, also known as AIM, which included a feature whereby users could “warn” other users, attaching a cost to, and thus disincentivizing, certain kinds of speech. The nuances of the design of this feature make it especially well-suited as a subject of study in value-sensitive design as it seems to be the product of significant reflection and foresight by its designers. In the course of examining this case, this paper proposes two novel and generalizable processes of integrating values into the design of technology, inspired by the approach of value-sensitive design: a “method of decomposition,” reconstructing a user journey in order to identify possible moments of intervention; and an iterative “Innovation–Abuse–Innovation” branching diagram, which systematizes the process of anticipating abuse cases and designing responses to them. These methods build upon recent work in the literature on operationalizing ethical values in the design process. I close by illustrating the flexibility and generalizability of these methods and speculating on how they might be applied to contemporary platforms.

This “literature refresh” identifies the most relevant new research in AI and robotic systems ethics from January 1, 2022 to January 31, 2024. Our selection methodology consisted of traditional research methods as well as novel human-AI teaming techniques, leveraging the expert human judgment of the authors, enhanced with a collection of AI and computational tools. We have identified stable trends in the critiques of the use of AI in the defense and security domain that cluster around worries about machine bias as well as the propensity of the technology to exacerbate human cognitive biases. Training data ambiguities, irregularities or untrustworthy data, and outright hacking of training sets are notable problems reported by the papers in our research set. This limits the trustworthiness of some systems which is heightened by the “black box” nature of many of these technologies which makes accountability and testing difficult. Given the speed and immense scale of operations that AI systems are involved in, there is a pronounced drift away from the reliance on “human in the loop” and “human on the loop” as the gold standard. We are now at the stage where a new ethical paradigm is needed.

We investigate the question of whether (and if so why) creating or distributing deepfake pornography of someone without their consent is inherently objectionable. We argue that nonconsensually distributing deepfake pornography of a living person on the internet is inherently pro tanto wrong in virtue of the fact that nonconsensually distributing intentionally non-veridical representations about someone violates their right that their social identity not be tampered with, a right which is grounded in their interest in being able to exercise autonomy over their social relations with others. We go on to suggest that nonconsensual deepfakes are especially worrisome in connection with this right because they have a high degree of phenomenal immediacy, a property which corresponds inversely to the ease with which a representation can be doubted. We then suggest that nonconsensually creating and privately consuming deepfake pornography is worrisome but may not be inherently pro tanto wrong. Finally, we discuss the special issue of whether nonconsensually distributing deepfake pornography of a deceased person is inherently objectionable. We argue that the answer depends on how long it has been since the person died.

Few sectors are more affected by COVID-19 than higher education. There is growing recognition that reopening the densely populated communities of higher education will require surveillance technologies, but many of these technologies pose threats to the privacy of the very students, faculty, and staff they are meant to protect. The authors have a history of working with our institution’s governing bodies to provide ethical guidance on the use of technologies, especially including those with significant implications for privacy. Here, we draw on that experience to provide guidelines for using surveillance technologies to reopen college campuses safely and responsibly, even under the specter of covid. We aim to generalize our recommendations, so they are sensitive to the practical realities and constraints that universities face.

To many, the idea of autonomous weapons systems (AWS) killing human beings is grotesque. Yet critics have had difficulty explaining why it should make a significant moral difference if a human combatant is killed by an AWS as opposed to being killed by a human combatant. The purpose of this paper is to explore the roots of various deontological concerns with AWS and to consider whether these concerns are distinct from any concerns that also apply to long-distance, human-guided weaponry. We suggest that at least one major driver of the intuitive moral aversion to lethal AWS is that their use disrespects their human targets by violating the martial contract between human combatants. On our understanding of this doctrine, service personnel cede a right not to be directly targeted with lethal violence to other human agents alone. Artificial agents, of which AWS are one example, cannot understand the value of human life. A human combatant cannot transfer his privileges of targeting enemy combatants to a robot. Therefore, the human duty-holder who deploys AWS breaches the martial contract between human combatants and disrespects the targeted combatants. We consider whether this novel deontological objection to AWS forms the foundation of several other popular yet imperfect deontological objections to AWS.

In July 2020, more than 1000 companies that advertise on social media platforms withdrew their business, citing failures of the platforms (especially Facebook) to address the proliferation of harmful content. The #StopHateForProfit movement invites reflection on an understudied topic: the ethics of boycotting by corporations. Under what conditions is corporate boycotting permissible, required, supererogatory, or forbidden? Although value-driven consumerism has generated significant recent discussion in applied ethics, that discussion has focused almost exclusively on the consumption choices of individuals. As this article underscores, value-driven consumerism by business corporations complicates these issues and invites further examination. We propose principles for the ethics of boycotting by corporations, indicate how these principles relate to different CSR paradigms, and show how these insights can help assess recent instances of corporate boycotting.

In this paper, we outline a new method for evaluating the human impact of machine-learning applications. In partnership with Underwriters Laboratories Inc., we have developed a framework to evaluate the impacts of a particular use of machine learning that is based on the goals and values of the domain in which that application is deployed. By examining the use of artificial intelligence in particular domains, such as journalism, criminal justice, or law, we can develop more nuanced and practically relevant understandings of key ethical guidelines for artificial intelligence. By decoupling the extraction of the facts of the matter from the evaluation of the impact of the resulting systems, we create a framework for the process of assessing impact that has two distinctly different phases.

"A runaway trolley is speeding down a track" So begins what is perhaps the most fecund thought experiment of the past several decades since its invention by Philippa Foot. Since then, moral philosophers have applied the "trolley problem" as a thought experiment to study many different ethical conflicts - and chief among them is the programming of autonomous vehicles. Nowadays, however, very few philosophers accept that the trolley problem is a perfect analogy for driverless cars or that the situations autonomous vehicles face will resemble the forced choice of the unlucky bystander in the original thought experiment. This book represents a substantial and purposeful effort to move the academic discussion beyond the trolley problem to the broader ethical, legal, and social implications that autonomous vehicles present. There are still urgent questions waiting to be addressed, for example: how AVs might interact with human drivers in mixed or "hybrid" traffic environments; how AVs might reshape our urban landscapes; what unique security or privacy concerns are raised by AVs as connected devices in the "Internet of Things"; how the benefits and burdens of this new technology, including mobility, traffic congestion, and pollution, will be distributed throughout society; and more. An attempt to map the landscape of these next-generation questions and to suggest preliminary answers, this volume draws on the disciplines of philosophy, sociology, economics, urban planning and transportation engineering, business ethics and more, and represents a global range of perspectives.

Rule consequentialism is usually taken to recommend a single ideal code for all moral agents. Here I argue that, depending on their theoretical mo- tivations, some rule consequentialists have good reasons to be relativists. Rule consequentialists who are moved by consequentialist considerations ought to support a scheme of multiple relativized moral codes because we could expect such a scheme to have better consequences in terms of impartial aggregate well- being than a single universal code. Rule consequentialists who nd compelling the theory’s coherence with our considered moral intuitions should do the same because a scheme of multiple codes could better cohere with our intuitions about costless bene ts, though these intuitions must be weighed against our allegiance to moral universalism.

Robert Sparrow argues that several initially plausible arguments in favor of the deployment of autonomous weapons systems (AWS) in warfare fail, and that their deployment faces a serious moral objection: deploying AWS fails to express the respect for the casualties of war that morality requires. We critically discuss Sparrow’s argument from respect and respond on behalf of some objections he considers. Sparrow’s argument against AWS relies on the claim that they are distinct from accepted weapons of war in that they either fail to transmit an attitude of respect or they transmit an attitude of disrespect. We argue that this distinction between AWS and widely accepted weapons is illusory, and so cannot ground a moral difference between AWS and existing methods of waging war. We also suggest that, if deploying conventional soldiers in some situation would be permissible, and if we could expect deploying AWS to cause fewer civilian casualties, then it would be consistent with an intuitive understanding of respect to deploy AWS in this situation.

We propose that the prevalent moral aversion to AWS is supported by a pair of compelling objections. First, we argue that even a sophisticated robot is not the kind of thing that is capable of replicating human moral judgment. This conclusion follows if human moral judgment is not codifiable, i.e., it cannot be captured by a list of rules. Moral judgment requires either the ability to engage in wide reflective equilibrium, the ability to perceive certain facts as moral considerations, moral imagination, or the ability to have moral experiences with a particular phenomenological character. Robots cannot in principle possess these abilities, so robots cannot in principle replicate human moral judgment. If robots cannot in principle replicate human moral judgment then it is morally problematic to deploy AWS with that aim in mind. Second, we then argue that even if it is possible for a sufficiently sophisticated robot to make ‘moral decisions’ that are extensionally indistinguishable from (or better than) human moral decisions, these ‘decisions’ could not be made for the right reasons. This means that the ‘moral decisions’ made by AWS are bound to be morally deficient in at least one respect even if they are extensionally indistinguishable from human ones. Our objections to AWS support the prevalent aversion to the employment of AWS in war. They also enjoy several significant advantages over the most common objections to AWS in the literature.

Many social trends are conspiring to drive the adoption of greater automation in society, and we will certainly see a greater offloading of human decisionmaking to robots in the future. Many of these decisions are morally salient, including decisions about how benefits and burdens are distributed. Roboticists and ethicists have begun to think carefully about the moral decision making apparatus for machines. Their concerns often center around the plausible claim that robots will lack many of the mental capacities that are indispensable in human moral decision making, such as empathy. To the extent that robots may be robustly artificially intelligent, these concerns subside, but they give way to new worries about creating artificial agents to do our bidding, if those artificial agents have moral standing. We suggest that the question of AI consciousness poses a dilemma. Whether artificially intelligent agents will be conscious or not, we will face serious difficulties in programming them to reliably make moral decisions.

ABSTRACTThe jus ad bellum criterion of right intention is a central guiding principle of just war theory. It asserts that a country’s resort to war is just only if that country resorts to war for the right reasons. However, there is significant confusion, and little consensus, about how to specify the CRI. We seek to clear up this confusion by evaluating several distinct ways of understanding the criterion. On one understanding, a state’s resort to war is just only if it plans to adhere to the principles of just war while achieving its just cause. We argue that the first understanding makes the CRI superfluous, because it can be subsumed under the probability of success criterion. On a second understanding, a resort to war is just only if a state’s motives, which explain its resort to war, are of the right kind. We argue that this second understanding of the CRI makes it a significant further obstacle to justifying war. However, this second understanding faces a possible infinite regress problem, wh...

Cyberweapons are software and software, at least intuitively, is nonphysical. Several authors have noted that this potentially renders problematic the application of normative frameworks like UN Charter Article 2(4) to cyberweapons. If Article 2(4) only proscribes the use of physical force, and if cyberweapons are nonphysical, then cyberweapons fall outside the purview of Article 2(4). This article explores the physicality of software, examining Stuxnet in particular. First, I show that with a few relatively uncontroversial metaphysical claims we can secure the conclusion that Stuxnet is physical. In particular, there exist instances of Stuxnet that are both located in space and causally efficacious, and this is very strong evidence for their being physical. Second, I argue that the question of physicality is actually irrelevant for the moral evaluation of an attack like Stuxnet because of its undeniably physical effects. Finally, I argue that some features of Stuxnet should make us optimistic about the prospects for discrimination and proportionality in cyberwarfare.

Desert is a notion ubiquitous in our moral discourse, and the importance of its dictates is perhaps clearest when dealing with the distribution of material resources. George Sher has provided one account of desert in wages, answering the question, “How do workers deserve their wage?” Sher relies on the violation of preexisting “independent standards” that dictate how much of a certain good we think people are entitled to in general. When these standards are violated, they call for an offsetting response at a later point in time in order to restore the moral equilibrium. I argue that this formalization of desert is flawed at the theoretical level and that it has further difficulties when applied to wages in particular. Lastly, I offer some brief remarks about what I think are the criteria for establishing desert in wages.