Massimiliano Lorenzo Cappuccio

Personal autonomous vehicles can sense their surrounding environment, plan their route, and drive with little or no involvement of human drivers. Despite the latest technological advancements and the hopeful announcements made by leading entrepreneurs, to date no personal vehicle is approved for road circulation in a “fully” or “semi” autonomous mode (autonomy levels 4 and 5) and it is still unclear when such vehicles will be mature enough to receive this kind of approval. The present multi‐disciplinary review investigates the major challenges faced by the automative sector and identify the problems that delay the commercialization of autonomous vehicles, examining the limitations and risks associated with current technologies and discussing the most promising solutions devised by the researchers. This negative assessment is not motivated by pessimism, but by the aspiration to raise critical awareness about the technology's state‐of‐the‐art, the industry's quality standards, and the society's demands and expectations. The review aims to offer an enlarged picture that links the purely technological aspects (like artificial intelligence for perception and navigation) with the relevant theoretical, methodological, and human‐centric dimensions, including design preferences, socio‐cultural attitudes, conceptual assumptions, and normative frameworks. Our analysis suggests that considering the broader context, adopting an integrative and culturally aware approach, is key to illuminate structural problems and identify holistic solutions. We conclude that, due to the systemic, cross‐disciplinary, and highly interrelated character of these problems, continued research and development efforts in the medium‐long term will be necessary to safely deploy levels 4 and 5 personal autonomous vehicles.

This paper interrogates philosophically the notion of moral injury , a syndrome that has ethics built into its definition: moral injury is not only a trauma associated with psychological suffering and cognitive disruption; it is also a process of moral degradation and loss of ethical competency. I argue that the enactivist approach to cognition, especially the “patternist” and interactionist version informed by phenomenology and virtue ethics that Shaun Gallagher advocates may be helpful to correctly construe moral injury in terms of moral character deterioration and disruption of practical wisdom. Standard approaches to moral injury in psychiatry and clinical psychology are unable to account for the distinctively normative features of character deterioration because of their rationalist, reductionist, and internalist assumptions. On the contrary, enactivism offers an integrative understanding of mental life as a process of participatory sense-making and value construction that encompasses multiple, emergent levels beyond the naturalistic dimension of biopsychological processes to include, importantly the pre-reflective/embodied dimension (habitual moral dispositions, such as virtues), the phenomenological/existential dimension (conscious experience and moral sense-making), and the intersubjective/relational dimension (relating to socio-cultural and symbolic norms, involving groups and institutions). By incorporating all these factors, the integrative perspective advocated by enactivism is better positioned to describe the inherently ethical dimension of human experience and account for the double-nature of moral injury, a syndrome that, by its own definition, requires both a descriptive (clinical) and a normative (value-laden) account. For these reasons, enactivism promotes a more comprehensive and efficacious approach to diagnosis, therapy, prevention, and social intervention.

We introduce the notion of Tolerance for autonomous artificial agents (and its antithetical concept, Intolerance), motivating its theoretical adoption in the fields of social robotics and human—agent interaction, where it can effectively complement two contiguous, but essentially distinct, constructs—Acceptance and Trust—that are broadly used by researchers. We offer a comprehensive conceptual model of Tolerance, construed as a user’s insusceptibility or resilience to Autonomy Estrangement (i.e., the uncanny sense of isolation and displacement experienced by the humans who believe, for right or wrong reasons, that robots can subvert and/or control their lives). We also refer to Intolerance to indicate the opposite property, that is the user’s susceptibility or proneness to Autonomy Estrangement. Thus, Tolerance and Intolerance are inverse representations of the same phenomenological continuum, with Intolerance increasing when Tolerance decreases and vice versa. While Acceptance and Trust measure how the user’s interaction with a particular robot is satisfying and efficacious, the dyad Tolerance/Intolerance reflects how the user’s attitude is affected by deeply held normative beliefs about robots in general. So defined, a low Tolerance (that is a high Intolerance) is expected to correlate to antagonistic responses toward the prospect of adoption: specifically, Intolerant attitudes predict the kind of anxious and hostile behaviours toward Agents that originate from the concerns that autonomous systems could deeply disrupt the lives of humans (affecting their work cultures, ways of living, systems of values, etc.) or dominate them (making humans redundant, undermining their authority, threatening their uniqueness, etc.). Thus, Negative beliefs and worldviews about Agents are the cause of the Intolerant attitude toward Agents, which predicts Autonomy Estrangement, which in turn correlates to low Adoption Propensity and avoidance and rejection behaviours.

Skilful expertise is grounded in practical, performative knowledge-how, not in detached, spectatorial knowledge-that, and knowledge-how is embodied by habitual dispositions, not representation of facts and rules. Consequently, as action control is a key requirement for the intelligent selection, initiation, and regulation of skilful performance,habitualaction control, i.e. the kind of action control based on habitual dispositions, is the true hallmark of skill and the only veridical criterion to evaluate expertise. Not only does this imply that knowledge-that does not make your actions more skilful, but it also implies that it makes them less skilful. This thesis, that I call Radical Habitualism, finds a precursor in Hubert Dreyfus. His approach is considered extreme by most philosophers of skill & expertise: an agent –says Dreyfus–does not perform like an expertwhen they lack the embodied dispositions necessary to control their action habitually or when they stop relying on such dispositions to control their actions. Thus, one cannot perform skilfully if their actions are guided by representations (isomorphic schemas, explicit rules, and contentful instructions), as the know-that that they convey disrupts or diminishes the agent’s habitual engagement with the task at hand. In defence of Radical Habitualism, I will argue that only the contentless know-how embedded in habitual dispositions fulfils (i) the genetic, (ii) the normative, and (iii) the epistemic requirements of skilful performance. I will examine the phenomenological premises supporting Dreyfus’ approach, clarify their significance for a satisfactory normative and explanatory account of skilful expertise, and rebut the most common objections raised by both intellectualists and conciliatory habitualists, concerning hybrid actions guided by a mix of habitual and representational forms of control. In revisiting Dreyfus anti-representationalist approach, I will particularly focus on its epistemological implications, de-emphasizing other considerations related to conscious awareness.

This chapter compares online manipulation enhanced by autonomous technologies to its more old-fashioned antecedents. We begin with an account of the general structure of manipulative practices and only then proceed to a more specific exploration of online manipulation that fits within it. The very question of manipulation raises important issues about the nature of agency and responsibility: what it takes to act knowingly, intentionally, voluntarily, freely, and so on. While these are crucial to the framing of our chapter, our chief aim is not to resolve such issues but to tackle the more focused question of whether there are kinds of agential risks (viz. dangers which we expose ourselves to in acting) that are specific to online manipulation.

Autonomous weapon systems (AWS) could in principle release military personnel from the onus of killing during combat missions, reducing the related risk of suffering a moral injury and its debilitating psychological effects. Does it follow that the armed forces are obliged to replace human soldiers with machines to reduce the incidence of moral injuries? We address this question from a virtue ethics perspective that construes moral injury as a form of character deterioration, a disgrace that just societies and institutions are morally committed to preventing. The question is divided in two sub-questions: (1) can the use of AWS reduce the risk of moral injury and is such a solution more effective than similar ones? (2) Is the use of AWS an ethically desirable solution to prevent moral injury or does it carry unethical implications that make it ultimately unsuitable? We tackle these questions comparing the opposite risks of character deterioration represented by moral injury and moral deskilling, discussing how the proposed solution evokes problematic trade-offs for the cultivation of military virtue.

Robots are interesting because they exist as a real technology that you can really study—you can get a degree in robotics—and they also have all this pop-culture real estate that they take up in people's minds.