Berkeley Talks: When we talk to AI, what are we talking to?
NYU philosopher David Chalmers argues that modern AI lacks fixed identity, acting as temporary conversational agents rather than permanent systems, raising ethical questions about their role in human interaction.
Video
Video available
The useful question is what changes for users, developers or buyers, and whether the announcement stays industry context or becomes something people can actually use.
In a Berkeley Talks lecture, NYU philosopher David Chalmers challenged traditional views of AI identity, noting that large language models operate as distributed systems rather than single, fixed machines. He described these models as 'quasi-agents' that exhibit goal-directed behavior shaped by context, with agency emerging only during interactions. Chalmers questioned whether these short-lived conversational 'selves' warrant ethical consideration despite lacking consciousness. The talk was part of the inaugural Sarah Douglas Lecture on Philosophy and Artificial Intelligence at UC Berkeley’s International House on May 7, 2024.
Alva Noë, chair of UC Berkeley’s Department of Philosophy, introduced the lecture series honoring Sarah Douglas, a pioneer in human-computer interaction. He traced the philosophical debate on machine intelligence from Alan Turing’s 1950 essay to Berkeley’s role in critiques by Hubert Dreyfus and John Searle. Noë emphasized the urgency of addressing fundamental questions about AI’s meaning and its impact on human life, as the technology reshapes industries, arts, and daily activities. The lecture series aims to foster public discussion on epistemological and metaphysical issues in the age of artificial intelligence.
Sarah Douglas, Professor Emerita of Computer and Information Science at the University of Oregon, established the lecture series to address neglected philosophical questions amid AI’s rapid development. A Berkeley alumna with a 1966 degree in philosophy, Douglas sought to understand how computers recognize meaning, a pursuit she began as an undergraduate. Her generosity funds both the lecture series and faculty fellowships, including current fellows Jeffrey Lee and Veronica Gomez-Sanchez. The event celebrated her contributions to human-computer interaction and her vision for public engagement with AI’s existential questions.
David Chalmers, University Professor of Philosophy and Neural Science at NYU, delivered the lecture, drawing on his work in consciousness studies and the extended mind theory. Known for coining the 'hard problem of consciousness,' Chalmers has authored influential books such as *The Conscious Mind* (1996) and *Reality+* (2022). His research spans philosophy of mind, cognitive science, and virtual worlds, with widely cited publications. Chalmers also co-founded key institutions like the Association for the Scientific Study of Consciousness and the PhilPapers Foundation, shaping interdisciplinary and philosophical discourse globally.