Events

Abstract: Over the past decade, AI research has focused heavily on building ever-larger deep learning models. This approach has simultaneously unlocked incredible achievements in science and technology, and hindered AI from overcoming long-standing limitations with respect to explainability, ethical harms, and environmental efficiency. Drawing on qualitative interviews and computational analyses, our three-part history of AI research traces the creation of this "epistemic monoculture" back to a radical reconceptualization of scientific progress that began in the 1980s. Amid an "AI Winter," an intervention by the U.S. government reoriented the field towards quantifiable progress on tasks of military and commercial interest through a new evaluation system called “benchmarking.” By distilling science down to verifiable metrics, benchmarking clarified the roles of scientists, allowed the field to rapidly integrate talent, and provided clear signals of significance and progress. But history has also revealed a tradeoff to this streamlined approach to science: the consolidation around external interests and inherent conservatism of benchmarking disincentivized exploration beyond scaling monoculture. In the discussion, we explain how AI's monoculture offers a compelling challenge to the belief that basic, exploration-driven research is needed for scientific progress. Implications for the spread of AI monoculture to other sciences in the era of generative AI are also discussed.

The Sociology Seminar Series for Trinity Term is convened by Juliana de Castro Galvao, Pablo Geraldo and David Kretschmer.  For more information about this or any of the seminars in the series, please contact sociology.secretary@nuffield.ox.ac.uk.