ABSTRACT
This chapter introduces a critical framework for understanding AI-generated images as sociotechnical assemblages rather than isolated visual outputs. It argues that visual generative AI emerges from deep infrastructures: massive training datasets, model architectures, commercial platforms, interface affordances, and the everyday practices through which images circulate as networked data. Rather than a radical rupture, AI images extend existing cultural techniques of image production, datafication, and representation while reshaping notions of agency and creativity. To grasp this transformation, the chapter proposes Six Critical Lenses organized around three analytical “realities”: the sociotechnical fabrics enabling generative systems; the semiotic interfaces that structure human–machine co-production; and the representational, aesthetic and political dimensions visible in AI images themselves. This framework connects machine operations to cultural meaning-making, highlighting the embedded ideologies and power geometries of training data, interface design, and emerging visualities.
