A semiotic analysis of the AI-driven reproduction of the superhero myth in marvel film posters

This study examined how the superhero myth was transformed when Marvel film posters were reproduced through generative artificial intelligence and to what extent character-specific heroic codes were preserved. The study is significant in demonstrating that AI-assisted visual production functioned not merely as a technical reproduction process but also as a field of meaning-making that reorganized mythological, ideological, and aesthetic codes embedded in popular culture images. This study was designed as a qualitative visual analysis, the research compared three original film posters featuring Iron Man, Captain America, and Thor with three AI-generated posters produced using the Gemini Nano Banana 2 model. The visuals were analysed through Roland Barthes’s concepts of denotation, connotation, and myth, focusing on character positioning, body posture, gaze, colour, lighting, costume, symbolic objects, background, and overall atmosphere. These categories were selected because they represent the main visual components through which superhero posters construct character identity, heroic meaning, and narrative atmosphere. The findings showed that artificial intelligence did not eliminate technological, national-military, and mythological heroic codes; instead, it reconstructed them through more explicit, symmetrical, intensified, and standardised visual patterns. Consequently, character-specific identities were partly preserved but shifted towards more generalised and easily recognisable superhero archetypes.