After a short break over the winter holidays — and a few weeks dedicated to planning our activities for the new year — MSMC resumed its traditional Midweek Conversations series for 2026.
On Wednesday, January 14, we have been pleased to host Guido Anselmi, Associate Professor at the University of Catania and a long-standing friend of our academic community.
His presentation on generative AI fits closely with the School’s core themes, offering new methodological insights into an urgent and rapidly evolving area of research that still calls for robust analytical tools.
Rent, Style, and Automation: Generative AI in the Platformization of Culture
Guido Anselmi – Universoity of Catania
The platformization of society, how digital platforms reshape economic, social, and cultural life, has become a key topic in recent scholarship. Central to this process is the extraction of value through intermediation and the shaping of user behavior by enabling certain actions while restricting others. This dynamic is particularly evident in the production and circulation of cultural content. With the rise of generative AI (Gen AI)—platforms specifically designed to produce cultural objects—and the stagnation of traditional platforms like social media and search engines, we must reconsider how platform affordances influence cultural production. In earlier work (Caliandro & Anselmi 2021), we argued that platforms like Instagram enable a “memetic logic” by standardizing content creation through visual formulas (e.g., selfies, outfits) that act as a communicative grammar. Platform power lies in balancing standardization with enough variation to sustain discourse.
This paper examines how this dynamic evolves with Gen AI. We analyze over 2 million Stable Diffusion prompts (2022) and nearly 1 million Midjourney prompts (2023) using natural language processing, thematic clustering, and TF-IDF. Our contribution is twofold: first, we offer a research protocol to study visual Gen AI outputs; second, we provide empirical evidence of platformization at work in Gen AI. Our findings show that prompts cluster not by subject (e.g., cats, women) but around anchor keywords referencing styles or specific artists. This reveals Gen AI’s role in magnifying existing cultural trends and supports a broader theory of platformization. Furthermore, Gen AI platforms emerge as monopolistic actors that facilitate rent-seeking in the cultural industry, reinforcing the logic of platform capitalism.
