In this program, Jaron Lanier, Microsoft's prime unifying scientist, discusses a piece he published in The New Yorker (“There Is No AI”) about applying data dignity ideas to artificial intelligence. Lanier argues that large-model AI can be reconceived as a social collaboration by the people who provide data to the model in the form of text, images and other modalities. This is a figure/ground inversion of the usual conception of AI as being a participant or collaborator in its own right. Explanations of model results and behaviors would then center around the relative influence of specific inputs through a provenance calculation mechanism. This formulation suggests new and different strategies for long-term economics in the context of high-performance AI, as well as more concrete approaches to many safety, fairness and alignment questions.
This program is co-hosted with the UC Berkeley College of Computing, Data Science, and Society and the UC Berkeley Artificial Intelligence Research (BAIR) Lab. The CITRIS Research Exchange delivers fresh perspectives on information technology and society from distinguished academic, industry and civic leaders. Series: "Data Science Channel" [Science] [Show ID: 39326]

Ver todos los episodios