The Project

Self-knowledge poses problems central to philosophy of mind, epistemology, phenomenology, psychology, and ethics. Difficult questions also arise concerning its object: Is self-knowledge about a self or subject, or about an aspect of experience? What is the importance of self-knowledge to moral agency, responsibility, and decision-making, and what impact do various forms of moral and mental cultivation have on it? These questions are present in the world’s great philosophical traditions and inform empirical work in cognitive science. Recent advances in AI raise additional questions about whether artificial systems could exhibit capabilities that reflect or mirror those of human agents, specifically those involving self-awareness and metacognition.

Goals of the Institute

The chief intellectual objective of this NEH Summer Institute is to provide a forum for an intensive exploration of four core issues concerning the problem of self-knowledge for humans and artificial systems.

Outline

The motivation for extending our inquiry to include self-knowledge in humans and artificial systems comes in recognition of the impactful ways in which developments in AI are already shaping the conversation about the many dimensions of the problem.

Contemporary philosophy of mind and phenomenology provide one systematic framework. They are concerned with issues such as whether knowledge of our own mental states is epistemically secure, whether claims regarding states of mind bear a specific authority or presumption of truth, and whether empirical advances in the study of belief formation challenge traditional assumptions about the distinctive features of self-knowledge. Equally valuable frameworks of analysis are found in various streams of South and East Asian philosophy. Philosophers in these traditions have developed a sophisticated literature and conceptual apparatus that, in many ways, complements those of the Western traditions. But attitudes regarding the scope and reach of self-knowledge and the consequences of building AI systems that exhibit the full range of metacognitive and empathic skills vary both within and between these traditions.

Societies in the East, particularly Japan, often contrast with those in the West with regard to their stance toward technology in general and AI in particular. Efforts, in the case of the former, to realize a technology-integrated society through disruptive innovations that adopt a positive approach to AI and robot-human co-existence––as illustrated by, among other things, artificial life projects––reflect more optimistic attitudes regarding AI (e.g., autonomy, non-maleficence, sustainability, solidarity) than is typical of more cautionary Western attitudes. Considering the impactful ways in which developments in AI are shaping the conversation about the nature and limits of human consciousness, cognition, and intelligence, we think these philosophical frameworks stand to gain renewed relevance and greater interdisciplinary reach in clarifying what is distinctively human and what may be replicable—or even surpassable—by machines.

For instance, the metaphysical dimension of the problem begins with questions such as these: Does self-knowledge pertain to a self or sense of self or to impersonal cognitive systems and processes functionally organized to monitor, regulate, and direct information flow within the cognitive architecture? If the latter, is self-knowledge reducible to information processes in the brain, or is it rooted in deeper biological and social imperatives for staying alive and sustaining a coherent form of agency within a world of others? The epistemological dimension arises when we ask how we gain access to this self or sense of self, real or imaginary, and whether that access is immediate or mediated, fallible or infallible. Phenomenological questions arise when we ask about the role of self-knowledge in the constitution of experience: are we necessarily or implicitly self-aware when we are aware of something or someone? If so, in what way and how? Are these epistemic and experiential dimensions of awareness computable, or do they reflect qualitative and normative features that are non-computable? Moral questions arise when we wonder whether, or to what degree, awareness of ourselves as selves or subjects is tied to moral agency, to morally reactive attitudes such as shame, pride, and resentment, and to moral responsibility. In what ways does the tendency to ascribe intentional agency and human traits to AI systems complicate these moral questions?

The best approach to these problems is, therefore, interdisciplinary and cross-cultural, and scholars and scientists addressing these questions and teachers developing curricula in these areas must be aware of how they are addressed and the insights that emerge from addressing them in different traditions and disciplinary settings.

The NEH Summer Institute on Self-knowledge for Humans and Artificial Systems at UC Berkeley was conceived with exactly this aim in mind. The Institute will feature seminar-style sessions led by visiting speakers and the project directors as well as collaborative workshops and presentations that relate the Institute’s themes to the participants' own pedagogical or research projects. For details about visiting speakers, session topics, and planned activities, please consult the Institute website.