Labeling of synthetic content and the legal duty of transparency

limits of lawfulness in the commercial use of generative artificial intelligence

Abstract

This article analyzes the labeling of synthetic content generated or manipulated by generative artificial intelligence as a legal duty of transparency and its limits of lawfulness in commercial use. From a dogmatic-legal, regulatory, and consumer protection perspective, it examines the normative foundations of the obligation to identify artificially produced or substantially modified content — including deepfakes, synthetic avatars, voice clones, and virtual influencers —, with emphasis on personality rights, the duty of information, the prohibition of misleading advertising, and the civil liability of economic actors. The central hypothesis holds that labeling constitutes an autonomous legal duty of transparency, derivable from the Brazilian system of trust protection, objective good faith, personal data protection, and consumer law, but has limited function: the identification of the synthetic nature of the content does not, by itself, legitimize its circulation when there is unauthorized use of image, voice, or identity, unlawful processing of personal data, misleading advertising, or violation of personality rights. Labeling is a condition of transparency, not a sufficient condition of lawfulness. The methodology is dogmatic-deductive, supported by specialized bibliographic review, normative analysis of current Brazilian law, and examination of comparative regulatory parameters, especially the European AI Act and the Digital Services Act. It concludes that the lawfulness of synthetic content requires a transparency architecture composed of labeling, consent, data protection, consumer law, accountability, and multi-subject civil liability.