Legal Cartography of Synthetic Identity: Image, Voice, Data Doubles and Artificial Intelligence

Imatge, Veu, Dobles de Dades i Intel·ligència Artificial

Abstract

This article proposes understanding synthetic identity as one of the most problematic forms taken by generative AI when it enables the production of voices, images, gestures, performances or statements attributable to real persons who have neither created, consented to nor controlled them. To address this phenomenon, the article takes as its starting point the concept of the data double formulated by Haggerty and Ericson, and brings it into dialogue with Nissenbaum’s contextual integrity and Solove’s taxonomy of privacy. From this framework, it argues that generative AI intensifies personal data extractivism by transforming such data into artificial presences capable of speaking, appearing or acting in someone’s name. On that basis, the article offers a legal cartography that organizes the responses provided by the General Data Protection Regulation, Spain’s Organic Law 1/1982 on honour, privacy and image rights, the Spanish Intellectual Property Act, the Spanish Trademark Act and the Artificial Intelligence Act, according to the technical operation involved, the material used, the context of circulation and the harm produced. It concludes that synthetic identity does not necessarily require autonomous regulation, but rather a better normative assemblage and an understanding of privacy as a common good.