Detecting Is Not Understanding: Functional, audiovisual, and legal analysis of the algorithmic mediation of human movement

Abstract: This paper examines how artificial intelligence mediates the detection and analysis of human movement in audiovisual contexts, particularly in film and advertising, treating movement as a complex object that cannot be reduced to technical measurability or expressive value. It shows how algorithmic translation affects the body, the meaning of the image, and the rights of the individuals involved, while optimization processes tend to normalize bodily gesture and automatic detection indirectly conditions aesthetic and narrative decisions. The study further situates bodily movement as information worthy of protection, raising questions about its capture, reuse, and circulation as data within automated systems. It argues that detecting movement is not equivalent to understanding it and concludes that preserving the plurality of interpretations of human movement requires treating artificial intelligence as a contextual tool, subordinated to narrative, functional, and normative frameworks. Note: Version updated in repository for preservation and digital accessibility. Original intellectual content remains unchanged. Generative AI tools were used solely for translation and linguistic refinement. No AI tools were involved in the conception or substantive development. The final text was reviewed and approved by the author. Official reference version:Zenodo DOI: https://doi.org/10.5281/zenodo.18463145Parallel version available:SSRN: https://ssrn.com/abstract=6168410 - DOI: https://doi.org/10.2139/ssrn.6168410Preservation copy at the Internet Archive: https://archive.org/details/zenodo.18463145