Gendered affective representations in AI-assisted music videos: a qualitative multimethod study

Generative artificial intelligence (AI) has become increasingly visible in music-video production, influencing both production practices and the ways in which emotion, bodies, and gender are portrayed visually. This study examines gendered emotional representation in AI-assisted music videos.

A qualitative multimethod design combined interpretive visual analysis of four purposively selected AI-assisted music videos with asynchronous written expert interviews involving six creative-industry professionals and four psychology experts (N = 10). The videos were analysed through an interpretive coding framework covering emotion, gender, body and face, power relations, AI aesthetics, and traces of bias.

Across the cases, emotion was conveyed not only through facial expression but also through atmosphere, colour, movement, editing, and spatial transitions. The selected works sometimes reproduced familiar conventions, including heteronormative intimacy, militarised masculinity, and idealised female bodies, but also unsettled them through female anger and solidarity and mutable synthetic bodies. Ambiguity in a body's appearance did not necessarily unsettle its gendered narrative role. Industry accounts described how prompting, selection, revision, and post-production shape representations, while general psychology-expert perspectives informed their interpretation.

The findings position AI-assisted music videos as co-produced sites of synthetic visual culture in which established gendered conventions may be repeated, negotiated, or challenged. Neither expert group provided direct evidence about the production or reception of the selected videos. The findings underline the practical importance of representational review, consent, diversity, and transparent disclosure in AI-assisted cultural production.