Senso comune artificiale. Egemonia culturale e intelligenza artificiale generativa

This article explores the role of Generative AI (GenAI) as an “epistemic gatekeeping” technology capable of influencing collective knowledge and individual epistemic agency. Through an experimental and qualitative approach, the study analyzes interactions between two Large Language Models (ChatGPT 4.0 and Claude 3.5 Sonnet) and ideal-typical users with different gender and political orientation (right/left) profiles, prompted on the controversial topic of surrogacy. The analysis focuses on the “genericity” of the outputs as a socio- technical code that produces an artificial common sense: a plausible and familiar discursive repertoire that mediates between content standardization and personalization. The findings reveal that while standardization is achieved through recurring discursive frames (legal, ethical, socio-economic), personalization adapts the content to the user’s profile, reinforcing stereotypes and crystallizing ideological positions. The article concludes that the primary risk of GenAI is not so much disinformation, but rather the progressive sedimentation of standardized discursive repertoires that, while appearing pluralistic, normalize a “stereotyped variety” and reinforce cultural hegemony by marginalizing non-mainstream positions.