The Universal Music Tokenizer: A Federated Architecture and DAO Governance Framework for Culturally Sovereign Music AI

Read full text Summarise with VIRAT
Authors: Philippe Astor
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

This whitepaper proposes the Universal Music Tokenizer (UMT), an open infrastructure initiative to address structural cultural bias in audio tokenization — the first and most determinative layer of any music AI pipeline. Drawing on quantitative evidence from CultureMERT (Kanatas et al., 2025), Papaioannou et al. (2025), and Nyandieka (2026), we document that dominant tokenizers systematically misrepresent non-Western musical traditions, producing Jensen-Shannon divergence scores of up to 22.6% between Western and non-Western corpora. We present four interconnected proposals: (1) a federated codec architecture using task arithmetic to merge community-trained cultural adapters without centralising raw data; (2) a multi-college DAO governance framework with structurally differentiated voting power and community veto rights; (3) a legal architecture grounded in OCAP® principles, Community Research Agreements, and a community copyright strategy for adapter weights; and (4) a Living Music Fund channelling 15% of commercial revenue into transmission, creation, and advocacy. Version 2.6. Working document — feedback welcome.

Fields of Study & Subjects

Ethics and Social Impacts of AI Music and Audio Processing Music Technology and Sound Studies

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
13 March 2026
Language
EN
Document type
Preprint