Tagliasacchi and Borsos are credited with constructing “Audio Overviews” for DeepMind, which developed an artificial intelligence feature that can turn text into a natural-flowing podcast.