Percy Liang, director of Stanford HAI’s Center for Research on Foundation Models, comments on the power and risks of large-scale AI models such as Google’s LaMDA.
