From Recording to Generating: The Reconstruction of Film Lens Aesthetics by Generative AI

Generative AI subverts the traditional film aesthetic paradigm based on photographic indexicality through algorithmic image synthesis, realizing a paradigmatic shift from "recording reality" to "generating reality" in film creation. Starting from the perspective of film ontology, this paper analyzes the technical logic of generative AI's image generation and its impact on the traditional "myth of total cinema". By exploring the changes in film lens language in composition, movement, light and shadow, and montage under the intervention of generative AI, it reveals the triple reconstruction logic of AI film aesthetics: the ontological shift from photographic ontology to algorithmic generation theory, the subject migration from single human creator to human-AI hybrid subject, and the formal fission of lens language driven by technological affordance. Combined with practical cases such as Sora and Runway Gen-2, this paper also reflects on the crises of hyperreality and emotional defocus in AI-generated images, and puts forward the development prospect of human-AI collaborative creation, providing a new theoretical perspective for film aesthetic research in the AI era.