Artificial Intelligence (AI) has rapidly become a culturally salient object, circulating in public discourse while also producing images and texts that shape social imaginaries. This chapter examines AI’s dual position in Social Representations Theory—as both Object and Alter in the semiotic triangle (Ego–Alter–Object)—by analysing human-authored textual descriptions of AI and AI-generated comics. We asked university students to write short stories about AI and to use them as prompts in an AI-based comic generator. We then conducted a thematic content analysis of the prompts and the resulting visuals (n = 110), focusing on anchoring and objectification processes. Findings reveal stable oppositions—AI as a tool vs. an autonomous entity; solution vs. risk—and a recurrent tension between disembodied intelligence and human-like embodiment. The visuals often anthropomorphise AI and, at the same time, reproduce gender and racial stereotypes embedded in training data and wider culture. The study shows how generative systems do not simply mirror participants’ narratives but recombine and amplify cultural meanings, contributing to the co-construction of social knowledge about AI. We discuss implications for public understanding and governance, including risks of overtrust through anthropomorphism and the persistence of biased representational patterns.
“Tell Me a Story About Yourself” Social Representations of Artificial Intelligence Through AI-Generated Comics
Rizzoli V.
;Sarrica M.
2026
Abstract
Artificial Intelligence (AI) has rapidly become a culturally salient object, circulating in public discourse while also producing images and texts that shape social imaginaries. This chapter examines AI’s dual position in Social Representations Theory—as both Object and Alter in the semiotic triangle (Ego–Alter–Object)—by analysing human-authored textual descriptions of AI and AI-generated comics. We asked university students to write short stories about AI and to use them as prompts in an AI-based comic generator. We then conducted a thematic content analysis of the prompts and the resulting visuals (n = 110), focusing on anchoring and objectification processes. Findings reveal stable oppositions—AI as a tool vs. an autonomous entity; solution vs. risk—and a recurrent tension between disembodied intelligence and human-like embodiment. The visuals often anthropomorphise AI and, at the same time, reproduce gender and racial stereotypes embedded in training data and wider culture. The study shows how generative systems do not simply mirror participants’ narratives but recombine and amplify cultural meanings, contributing to the co-construction of social knowledge about AI. We discuss implications for public understanding and governance, including risks of overtrust through anthropomorphism and the persistence of biased representational patterns.Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.




