The authors explore the interfaces between different dimensions of moral cognition, bridging economic, Bayesian and reinforcement learning perspectives. The human aversion to harming others cuts across these different interfaces, influencing decisions, judgments, and inferences about morality.

Scientific progress in recent neurofeedback research may bring about a new type of moral neuroenhancement, namely, neurofeedback-based moral enhancement; however, this has yet to be examined thoroughly. This paper presents an ethical analysis of the possibility of neurofeedback-based moral enhancement and demonstrates that this type of moral enhancement sheds new light on the moral enhancement debate. First, I survey this debate and extract the typical structural flow of its arguments. Second, by applying structure to the case of neurofeedback-based moral enhancement, (...) I examine the ethical, legal, and social issues to show that this technique is unique and traditionalist, which makes it compatible with almost all our conservative notions, so that it, accordingly, can be seen as an ethically acceptable option. Third, by rejecting the premise in the moral enhancement debate that bio/neuro-enhancement has its unique ELSI that traditional methods would never create, I demonstrate that, by virtue of its traditional or conservative features, neurofeedback-based moral enhancement can be incorporated into the traditional moral education network. Finally, I conclude that, being a part of the traditional moral education network, neurofeedback-based moral enhancement can be a unique and ethically acceptable option of moral neuroenhancement. (shrink)

The past 15 years occasioned an extraordinary blossoming of research into the cognitive and affective mechanisms that support moral judgment and behavior. This growth in our understanding of moral mechanisms overshadowed a crucial and complementary question, however: How are they learned? As this special issue of the journal Cognition attests, a new crop of research into moral learning has now firmly taken root. This new literature draws on recent advances in formal methods developed in other domains, such as Bayesian inference, (...) reinforcement learning and other machine learning techniques. Meanwhile, it also demonstrates how learning and deciding in a social domain—and especially in the moral domain—sometimes involves specialized cognitive systems. We review the contributions to this special issue and situate them within the broader contemporary literature. Our review focuses on how we learn moral values and moral rules, how we learn about personal moral character and relationships, and the philosophical implications of these emerging models. (shrink)

The fast pace of technological change necessitates new evaluative and deliberative tools. This article develops a general, functional approach to evaluating technological change, inspired by Nissenbaum’s theory of contextual integrity. Nissenbaum introduced the concept of contextual integrity to help analyze how technological changes can produce privacy problems. Reinterpreted, the concept of contextual integrity can aid our thinking about how technological changes affect the full range of human concerns and values—not only privacy. I propose a generalized concept of contextual integrity that (...) is applicable to a broader variety of circumstances, and I outline a new, general procedure for technological evaluation. Among the attractive features of the proposed approach to evaluating technological change are its context-sensitivity, adaptability, and principled presumptive conservatism, enabled by the mechanism the approach supplies for reevaluating existing practices, norms, and values. (shrink)

Research on ‘moral learning’ examines the roles of domain-general processes, such as Bayesian inference and reinforcement learning, in the development of moral beliefs and values. Alert to the power of these processes, and equipped with both the analytic resources of philosophy and the empirical methods of psychology, ‘moral learners’ are ideally placed to discover the contributions of nature, nurture and culture to moral development. However, I argue that to achieve these objectives research on moral learning needs to overcome nativist bias, (...) and distinguish two kinds of social learning: learning from and learning about. An agent learns from others when there is transfer of competence—what the learner learns is similar to, and causally dependent on, what the model knows. When an agent learns about the social world there is no transfer of competence—observable features of other agents are just the content of what-is-learned. Even learning from does not require explicit instruction. A novice can learn from an expert who is ‘leaking’ her morality in the form of emotionally charged behaviour or involuntary use of vocabulary. To the extent that moral development depends on learning from other agents, there is the potential for cultural selection of moral beliefs and values. (shrink)

Group representations based on recursive utilities can be used to derive the same predictions as Pietraszewski in conflict situations. Additionally, these representations generalize to non-conflict situations, asymmetric relationships, and represent the stakes in a conflict. However, both proposals fail to represent asymmetries of power and responsibility and to account for generalizations from specific observed individuals to collections of non-observed individuals.