Abstract

Artificial intelligence (AI) has become a structural element of contemporary cultural systems, influencing how meaning, language, and identity are created and circulated. Its pervasive role raises a set of intertwined problems that affect the preservation of cultural diversity and the visibility of non-dominant traditions. The rapid expansion of AI-generated content reinforces cultural homogenization and the invisibility of minority voices. The structure of algorithms and the biases embedded in training data reproduce hegemonic narratives, while the centralization of digital platforms concentrates control over access to cultural and creative content. This process generates new forms of discrimination, under-representation, and dependency within global cultural flows.

Further issues arise from over-reliance on synthetic data, leading to model collapse and loss of epistemic diversity, and from the absence of cultural impact assessments in existing regulatory frameworks. The EU AI Act, for instance, remains largely sector-neutral, neglecting the cultural implications of AI systems.

The paper examines these challenges—ethical, legal, and epistemic—that threaten cultural pluralism in the age of artificial intelligence.

TABLE OF CONTENTS

1. UNESCO’s Cultural Policy Vision and AI in Culture

2. The Role of Cultural Data in AI Systems
   2.1 The Data Value Divide
   2.2 From Recognition to Regulation

3. Risks for Cultural Diversity
   3.1 Bias and Hegemony
   3.2 Model Collapse and Cognitive Erosion
   3.3 Linguistic and cultural diversity

4. Cultural Rights within AI Governance
   4.1 The Scope of the CoE Framework Convention
   4.2 Human Oversight and Cultural Dignity

5. Policy and Regulatory Gaps
   5.1 The Limits of Current Frameworks

6. Conclusion