Robert Sparrow

Social isolation and loneliness are growing concerns around the globe that put people at increased risk of disease and early death. One much‐touted approach to addressing them is deploying artificially intelligent agents to serve as companions for socially isolated and lonely people. Focusing on digital humans, we consider evidence and ethical arguments for and against this approach. We set forth and defend public health policies that respond to concerns about replacing humans, establishing inferior relationships, algorithmic bias, distributive justice, and data privacy.

It is estimated that one in three women experience intimate partner violence (IPV) across the course of their life. The popular uptake of “smart speakers” powered by sophisticated AI means that surveillance of the domestic environment is increasingly possible. Correspondingly, there are various proposals to use smart speakers to detect or report IPV. In this paper, we clarify what might be possible when it comes to combatting IPV using existing or near-term technology and also begin the project of evaluating this project both ethically and politically. We argue that the ethical landscape looks different depending on whether one is considering the decision to develop the technology or the decision to use it once it has been developed. If activists and governments wish to avoid the privatisation of responses to IPV, ubiquitous surveillance of domestic spaces, increasing the risk posed to members of minority communities by police responses to IPV, and the danger that more powerful smart speakers will be co-opted by men to control and abuse women, then they should resist the development of this technology rather than wait until these systems are developed. If it is judged that the moral urgency of IPV justifies exploring what might be possible by developing this technology, even in the face of these risks, then it will be imperative that victim-survivors from a range of demographics, as well as government and non-government stakeholders, are engaged in shaping this technology and the legislation and policies needed to regulate it.

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.

In ‘in vitro eugenics’ (IVE), I outlined a theoretical use of a technology of artificial gametogenesis, wherein repeated iterations of the derivation of gametes from embryonic stem cells, followed by the fusion of gametes to create new embryos, from which new stem cells could be derived, would allow researchers to create multiple generations of human embryos in the laboratory and also to produce ‘enhanced’ human beings with desired traits. As a number of commentators observed, my purpose in publishing this paper was to provoke ethical discussion of a largely unremarked upon technological possibility and surrounding issues. Even if this was, as Murphy observes archly, to aim ‘low’, discussion of IVE is valuable for three reasons. First, it may render us better prepared should IVE become practical. I noted of my original discussion that it was speculative and several of the respondents suggest that IVE is even less likely to come about than I allowed there. Nevertheless, second, discussion of IVE is valuable for what it reveals about the ethics of new reproductive technologies (NRTs) more generally and, third, about the ethics of genetic human enhancement in particular. The responses to my paper demonstrate this nicely by (a) illustrating the selective way in which arguments about risk are mobilised in debates about NRTs and (b) highlighting the tension between any obligation of ‘procreative beneficence’ and a concern for genetic relatedness. Even if IVE should never be pursued, then, discussion of this possibility may help us better understand the ethics of other NRTs and means of genetic human enhancement.

A series of recent scientific results suggest that, in the not-too-distant future, it will be possible to create viable human gametes from human stem cells. This paper discusses the potential of this technology to make possible what I call ‘in vitro eugenics’: the deliberate breeding of human beings in vitro by fusing sperm and egg derived from different stem-cell lines to create an embryo and then deriving new gametes from stem cells derived from that embryo. Repeated iterations of this process would allow scientists to proceed through multiple human generations in the laboratory. In vitro eugenics might be used to study the heredity of genetic disorders and to produce cell lines of a desired character for medical applications. More controversially, it might also function as a powerful technology of ‘human enhancement’ by allowing researchers to use all the techniques of selective breeding to produce individuals with a desired genotype

That technologies may raise ethical issues is now widely recognised. The ‘responsible innovation’ literature – as well as, to a lesser extent, the applied ethics and bioethics literature – has responded to the need for ethical reflection on technologies by developing a number of tools and approaches to facilitate such reflection. Some of these instruments consist of lists of questions that people are encouraged to ask about technologies – a methodology known as the ‘Socratic approach’. However, to date, these instruments have often not adequately acknowledged various political impacts of technologies, which are, I suggest, essential to a proper account of the ethical issues they raise. New technologies can make some people richer and some people poorer, empower some and disempower others, have dramatic implications for relationships between different social groups and impact on social understandings and experiences that are central to the lives, and narratives, of denizens of technological societies. The distinctive contribution of this paper, then, is to offer a revised and updated version of the Socratic approach that highlights the political, as well as the more traditionally ethical, issues raised by the development of new technologies.

Artificial intelligence (AI) systems have demonstrated impressive performance across a variety of clinical tasks. However, notoriously, sometimes these systems are “black boxes.” The initial response in the literature was a demand for “explainable AI.” However, recently, several authors have suggested that making AI more explainable or “interpretable” is likely to be at the cost of the accuracy of these systems and that prioritizing interpretability in medical AI may constitute a “lethal prejudice.” In this paper, we defend the value of interpretability in the context of the use of AI in medicine. Clinicians may prefer interpretable systems over more accurate black boxes, which in turn is sufficient to give designers of AI reason to prefer more interpretable systems in order to ensure that AI is adopted and its benefits realized. Moreover, clinicians may be justified in this preference. Achieving the downstream benefits from AI is critically dependent on how the outputs of these systems are interpreted by physicians and patients. A preference for the use of highly accurate black box AI systems, over less accurate but more interpretable systems, may itself constitute a form of lethal prejudice that may diminish the benefits of AI to—and perhaps even harm—patients.

Objectives: Machine learning (ML) has the potential to facilitate “continual learning” in medicine, in which an ML system continues to evolve in response to exposure to new data over time, even after being deployed in a clinical setting. In this article, we provide a tutorial on the range of ethical issues raised by the use of such “adaptive” ML systems in medicine that have, thus far, been neglected in the literature. Target audience: The target audiences for this tutorial are the developers of ML AI systems, healthcare regulators, the broader medical informatics community, and practicing clinicians. Scope: Discussions of adaptive ML systems to date have overlooked the distinction between 2 sorts of variance that such systems may exhibit—diachronic evolution (change over time) and synchronic variation (difference between cotemporaneous instantiations of the algorithm at different sites)—and underestimated the significance of the latter. We highlight the challenges that diachronic evolution and synchronic variation present for the quality of patient care, informed consent, and equity, and discuss the complex ethical trade-offs involved in the design of such systems.

Artificial intelligence (AI) systems have demonstrated impressive performance across a variety of clinical tasks. However, notoriously, sometimes these systems are 'black boxes'. The initial response in the literature was a demand for 'explainable AI'. However, recently, several authors have suggested that making AI more explainable or 'interpretable' is likely to be at the cost of the accuracy of these systems and that prioritising interpretability in medical AI may constitute a 'lethal prejudice'. In this paper, we defend the value of interpretability in the context of the use of AI in medicine. Clinicians may prefer interpretable systems over more accurate black boxes, which in turn is sufficient to give designers of AI reason to prefer more interpretable systems in order to ensure that AI is adopted and its benefits realised. Moreover, clinicians may be justified in this preference. Achieving the downstream benefits from AI is critically dependent on how the outputs of these systems are interpreted by physicians and patients. A preference for the use of highly accurate black box AI systems, over less accurate but more interpretable systems , may itself constitute a form of lethal prejudice that may diminish the benefits of AI to - and perhaps even harm - patients.

ABSTRACT‘Rape by deception’ occurs when the victim ‘consents’ to sexual penetration as a result of certain sorts of deception by the perpetrator. The legal and philosophical literature on rape by deception has almost exclusively concentrated on cases wherein victims are brought to ‘consent’ to sexual intercourse by deception. Broadening our focus to consider sexual penetration in other contexts reveals a puzzle: if penetration in the context of sexual intercourse premised on deception is rape, is sexual penetration in the context of masturbation as a result of deception rape?

It is likely that sex robots will exist in the near future, making the effect they might have on human relationships a pressing concern. In this future world, we can imagine sex robots shaping our personal and social relationships through their unique access to, and potential for influencing, our most intimate of behaviours. We investigate whether they might be employed to influence social behaviours in a positive way. The paper begins with an account of the state of the art, acknowledges powerful feminist criticisms that have been made of sex robots, and evaluates suggestions that it might be possible to design sex robots which do not raise these concerns and which might even work to influence social behaviours in a positive way. It then outlines a number of ways that sex robots might be used to educate, “nudge”, and influence people in positive ways. It defends the idea that it would be ethical to use sex robots to promote socially positive behaviours — behaviours that benefit others and improve social cohesion, such as fostering respect and empathy for persons — but not to promote commercial products for parochial interests. We argue that the former project could advance individual and social welfare, while preserving personal autonomy — a minimum requirement of which is the ability to make informed decisions — whereas the latter depends on a lack of transparency and democratic control for its success, targets the vulnerability of the user to achieve its ends, and reinforces the problematic symbolism of negatively gendered sexuality. If sex robot design and application meet public requirements of transparency — enabling informed consent and reflective decision making — and democratic oversight — promoting accountability and the sharing of power with the public — it is conceivable that sex robots might assist, rather than harm society.

In order to avoid the implication that ‘mitochondrial replacement techniques’ (MRT) would produce ‘three parent babies’, discourses around these techniques typically dismiss the contribution of the mitochondria to genetic parenthood and personal identity. According to many participants in debates about MRT, ‘real parenthood’ is a matter of contributing nuclear DNA, which in turn implies that men and women make the same contribution to the embryo. Even when the importance of the mitochondria is acknowledged, an emphasis on mitochondrial DNA still has the effect of valorizing the role of DNA (and thus the paternal contribution to conception) at the expense of the role played by the cytoplasm of the oocyte in the development of the embryo and placenta, and that of the mother’s body in gestation. In this way, discourses around MRT falsely imply that what men and women contribute to reproduction and parenthood is the same—nuclear DNA—and thus erase the distinctive contribution that women make to conception. The potential of MRT to reconfigure relationships between the sexes in the service of patriarchal norms is perhaps one of the most significant things about it and should, we argue, be counted in the discussion of the ethical and policy implications of legitimating these procedures.

The fact that real-world decisions made by artificial intelligences (AI) are often ethically loaded has led a number of authorities to advocate the development of “moral machines”. I argue that the project of building “ethics” “into” machines presupposes a flawed understanding of the nature of ethics. Drawing on the work of the Australian philosopher, Raimond Gaita, I argue that ethical dilemmas are problems for particular people and not (just) problems for everyone who faces a similar situation. Moreover, the force of an ethical claim depends in part on the life history of the person who is making it. For both these reasons, machines could at best be engineered to provide a shallow simulacrum of ethics, which would have limited utility in confronting the ethical and policy dilemmas associated with AI.

What does Artificial Intelligence (AI) have to contribute to health care? And what should we be looking out for if we are worried about its risks? In this paper we offer a survey, and initial evaluation, of hopes and fears about the applications of artificial intelligence in medicine. AI clearly has enormous potential as a research tool, in genomics and public health especially, as well as a diagnostic aid. It’s also highly likely to impact on the organisational and business practices of healthcare systems in ways that are perhaps under-appreciated. Enthusiasts for AI have held out the prospect that it will free physicians up to spend more time attending to what really matters to them and their patients. We will argue that this claim depends upon implausible assumptions about the institutional and economic imperatives operating in contemporary healthcare settings. We will also highlight important concerns about privacy, surveillance, and bias in big data, as well as the risks of over trust in machines, the challenges of transparency, the deskilling of healthcare practitioners, the way AI reframes healthcare, and the implications of AI for the distribution of power in healthcare institutions. We will suggest that two questions, in particular, are deserving of further attention from philosophers and bioethicists. What does care look like when one is dealing with data as much as people? And, what weight should we give to the advice of machines in our own deliberations about medical decisions?

If people are inclined to attribute race to humanoid robots, as recent research suggests, then designers of social robots confront a difficult choice. Most existing social robots have white surfaces and are therefore, I suggest, likely to be perceived as White, exposing their designers to accusations of racism. However, manufacturing robots that would be perceived as Black, Brown, or Asian risks representing people of these races as slaves, especially given the historical associations between robots and slaves at the very origins of the project of robotics. The only way engineers might avoid this ethical and political dilemma is to design and manufacture robots to which people will struggle to attribute race. Doing so, however, would require rethinking the relationship between robots and “the social,” which sits at the heart of the project of social robotics. Discussion of the race politics of robots is also worthwhile because of the potential it has to generate insights about the politics of artifacts, the relationship between culture and technology, and the responsibilities of engineers.

Nancy Jecker is right when she says that older persons ought not to be ashamed if they wish to remain sexually active in advanced old age. She offers a useful account of the role that sexuality plays in supporting key human capabilities. However, Jecker assumes an exaggerated account of what sex robots are likely to be able to offer for the foreseeable future when she suggests that we are obligated to make them available to older persons with disabilities. Moreover, whether older persons should be ashamed to desire sex robots—or, more importantly, whether we should be ashamed at the thought that we should respond to the sexual needs of older persons by providing them with sex robots—turns on a range of arguments that Jecker fails to adequately consider.

In Deep Medicine, Eric Topol argues that the development of artificial intelligence (AI) for healthcare will lead to a dramatic shift in the culture and practice of medicine. Topol claims that, rather than replacing physicians, AI could function alongside of them in order to allow them to devote more of their time to face-to-face patient care. Unfortunately, these high hopes for AI-enhanced medicine fail to appreciate a number of factors that, we believe, suggest a radically different picture for the future of healthcare. Far from facilitating a return to “the golden age of doctoring”, the role of economic and institutional considerations in determining how medical AI will be used mean that it is likely to further erode therapeutic relationships and threaten professional and patient satisfaction.

Sex robots are likely to play an important role in shaping public understandings of sex and of relations between the sexes in the future. This paper contributes to the larger project of understanding how they will do so by examining the ethics of the “rape” of robots. I argue that the design of realistic female robots that could explicitly refuse consent to sex in order to facilitate rape fantasy would be unethical because sex with robots in these circumstances is a representation of the rape of a woman, which may increase the rate of rape, expresses disrespect for women, and demonstrates a significant character defect. Even when the intention is not to facilitate rape, the design of robots that can explicitly refuse consent is problematic due to the likelihood that some users will experiment with raping them. Designing robots that lack the capacity to explicitly refuse consent may be morally problematic depending on which of two accounts of the representational content of sex with realistic humanoid robots is correct.

Despite the advent of CRISPR, safe and effective gene editing for human enhancement remains well beyond our current technological capabilities. For the discussion about enhancing human beings to be worth having, then, we must assume that gene-editing technology will improve rapidly. However, rapid progress in the development and application of any technology comes at a price: obsolescence. If the genetic enhancements we can provide children get better and better each year, then the enhancements granted to children born in any given year will rapidly go out of date. Sooner or later, every modified child will find him- or herself to be “yesterday’s child.” The impacts of such obsolescence on our individual, social, and philosophical self-understanding constitute an underexplored set of considerations relevant to the ethics of genome editing.

There is increasing speculation within military and policy circles that the future of armed conflict is likely to include extensive deployment of robots designed to identify targets and destroy them without the direct oversight of a human operator. My aim in this paper is twofold. First, I will argue that the ethical case for allowing autonomous targeting, at least in specific restricted domains, is stronger than critics have acknowledged. Second, I will attempt to uncover, explicate, and defend the intuition that even in this context there would be something ethically problematic about such targeting. I argue that an account of the non-consequentialist foundations of the principle of distinction suggests that the use of autonomous weapon systems is unethical by virtue of failing to show appropriate respect for the humanity of our enemies. However, the success of the strongest form of this argument depends upon understanding the robot itself as doing the killing. To the extent that we believe that, on the contrary, AWS are only the means whereby those who order them into action kill, the idea that the use of AWS fails to respect the humanity of our enemy will turn upon an account of what is required by respect that is essentially conventional. Thus, while the theoretical foundations of the idea that AWS are weapons that are “evil in themselves” are weaker than critics have sometimes maintained, they are nonetheless sufficient to the task of demanding a prohibition of the development and deployment of such weapons.

In “Would we even know moral bioenhancement if we saw it?”, Harris Wiseman highlights a number of distinctions, between cognitive and emotional enhancement, voluntary and compulsory enhancement, and between enhancement and therapy, which he holds, not unreasonably, to be relevant to the debate about moral bioenhancement. He also offers a new distinction, between “hard” and soft moral bioenhancement, to which he believes critics of moral bioenhancement should be paying more attention. Having made these distinctions, Wiseman, suggests that it would be wrong to rely on any of them to try to settle the ethics of moral bioehancement and that instead we need to pay close attention to context in order to determine the ethics of any particular intervention. In this commentary, I will suggest that: (1) there is a further distinction to which we need to pay attention, which Wiseman neglects, between bioenhancement and other methods of shaping peoples’ behaviour and dispositions; and (2), that both the distinction between therapy and enhancement and Wiseman’s own distinction between hard and soft moral bioenhancement, are of more import than he allows when it comes to how we should respond to the burgeoning literature on moral bioenhancement.

In ‘Moral Enhancement, Freedom, and the God Machine’, Savulescu and Persson argue that recent scientific findings suggest that there is a realistic prospect of achieving ‘moral enhancement’ and respond to Harris's criticism that this would threaten individual freedom and autonomy. I argue that although some pharmaceutical and neuro‐scientific interventions may influence behaviour and emotions in ways that we may be inclined to evaluate positively, describing this as ‘moral enhancement’ presupposes a particular, contested account, of what it is to act morally and implies that entirely familiar drugs such as alcohol, ecstasy, and marijuana are also capable of making people ‘more moral’. Moreover, while Savulescu and Persson establish the theoretical possibility of using drugs to promote autonomy, the real threat posed to freedom by ‘moral bioenhancement’ is that the ‘enhancers’ will be wielding power over the ‘enhanced’. Drawing on Pettit's notion of ‘freedom as non‐domination’, I argue that individuals may be rendered unfree even by a hypothetical technology such as Savulescu and Persson's ‘God machine’, which would only intervene if they chose to act immorally. While it is impossible to rule out the theoretical possibility that moral enhancement might be all‐things‐considered justified even where it did threaten freedom and autonomy, I argue that any technology for biomedical shaping of behaviour and dispositions is much more likely to be used for ill rather than good.

In their paper, “Autonomy and the ethics of biological behaviour modification”, Savulescu, Douglas, and Persson discuss the ethics of a technology for improving moral motivation and behaviour that does not yet exist and will most likely never exist. At the heart of their argument sits the imagined case of a “moral technology” that magically prevents people from developing intentions to commit seriously immoral actions. It is not too much of a stretch, then, to characterise their paper as a thought experiment in service of a thought experiment. In order for an argument involving a thought experiment to progress debate in applied ethics three things must be true. First, the thought experiment must accurately represent and illuminate a pressing ethical dilemma. Second – and most obviously – the central claims of the argument regarding the thought experiment must be plausible. Third, it must be possible to apply or develop the arguments established with reference to the thought experiment to the real world cases the experiment is intended to illuminate. In this commentary, I argue that there are serious reasons to question the extent to which their argument meets each of these challenges involved in the use of thought experiments in applied ethics. While Savulescu et al. succeed in showing how behavioural modification might be compatible with freedom and autonomy – and perhaps justifiable even if it were not — in the fantastic case they consider, there is little we can conclude from this about any technology of “moral bioenhancement” in the foreseeable future. Indeed, there is a real danger that their argument will license attempts to manipulate behaviour through drugs and brain implants, which raise profound moral issues that they barely mention.