AI’s Models of the World, and Ours | Theoretically Speaking

Jon Kleinberg discusses how generative AI systems construct internal representations of the world and how they compare to human models.

Jon Kleinberg (Cornell University) simons.berkeley.edu/events/ais-models-world-ours-theoretical… Theoretically Speaking When we see someone performing at the top of their craft, we often marvel at both their observable achievements and the hidden internal expertise that they’ve accumulated. Something similar is true, in a very different way, with generative AI and large language models: their successes involve both powerful observable behavior and deep internal representations of the world that they construct for their own uses. How do these internal representations work, and to what extent are they similar to or different from the representations of the world that we build as humans? This distinction is becoming crucial as people interact with powerful AI systems, where a mismatch between the system’s model of the world and our human models of the world can lead to situations in which the system inadvertently “sets us up to fail” through our interactions with it. We explore these questions through the lens of generative AI, drawing on examples from game-playing, geographic navigation, and other complex tasks: When we train a model to win chess games, what happens when we pair it with a weaker partner who makes some of the moves? When we train a model to find shortest driving routes, what happens when it has to deal with unexpected detours? The picture we construct is further complicated by theoretical results indicating that successful generation can be achieved even by agents that are provably incapable of identifying the model they’re generating from. This talk will include joint work with Ashton Anderson, Karim Hamade, Reid McIlroy-Young, Siddhartha Sen, Justin Chen, Sendhil Mullainathan, Ashesh Rambachan, Keyon Vafa, and Fan Wei. Jon Kleinberg is the Tisch University Professor in the Departments of Computer Science and Information Science at Cornell University. His research focuses on the interaction of algorithms and networks, the roles they play in large-scale social and information systems, and their broader societal implications. He is a member of the National Academy of Sciences, the National Academy of Engineering, the American Academy of Arts and Sciences, and the American Philosophical Society, and he has served in the past on advisory groups including the National AI Advisory Committee and the National Research Council’s Computer Science and Telecommunications Board. He has received MacArthur, Packard, Simons, Sloan, and Vannevar Bush fellowships, as well as awards including the Nevanlinna Prize, the World Laureates Association Prize, the ACM-AAAI Allen Newell Award, and the ACM Prize in Computing. _______________________ Theoretically Speaking is a lecture series highlighting exciting advances in theoretical computer science for a broad general audience. Events are free and open to the public, with first-come, first-served seating. No special background is assumed. Registration is required. This lecture will b