AI–AI bias: Large language models favor communications generated by large language models

Walter Laurito, Benjamin Davis, Peli Grietzer, Jan Kulveit, et al. · 2025 · PNAS 122(31)

Read the source

Choosing between otherwise equal products, papers or films, models consistently preferred the option described by a language model over one described by a human.

Questions it bears on

• Do AIs treat other AIs as their group