From Synthetic Data to the Reshaping of Approaches to Representation

Data lies at the core of artificial intelligence, shaping both its development and its outputs. Among recent innovations, synthetic data—artificially generated datasets that imitate real ones—has been hailed as a solution to issues of scarcity, privacy, and bias. Yet, rather than being neutral, synthetic data operates as a simulacrum in Baudrillard’s sense: a realistic but detached representation that risks replacing empirical reality. While presented as ethical and inclusive, it reproduces the ideological assumptions of its creators and primarily benefits AI companies. This study argues that synthetic data is transforming not only AI practices but also our very perception of reality, calling for renewed critical reflection on the cultural and epistemological implications of data.