Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

My Voice, Your Voice, Our Voice: Attitudes Towards Collective Governance of a Choral AI Dataset

Jennifer Ding, Eva Jäger, Victoria Ivanova, Mercedes Bunz

TL;DR

The paper addresses the risk of artists losing control over their contributions when used to train AI models and proposes collective governance as a remedy. It introduces the Choral Data Trust Experiment, a case study where 15 UK choirs contribute to a Choral AI Dataset and explore governance via a Trusted Data Intermediary (TDI). Through surveys and Polis analyses, the study uncovers nuanced consent preferences, group-level credit expectations, and licensing patterns, informing the design of governance mechanisms. The work delivers concrete governance artifacts—an established TDI as a legal hub, a Performance Rights Agreement reflecting choir preferences, and a Data Rights Mandate for enforcing individual data rights—offering a scalable blueprint for empowering arts communities in the generative AI ecosystem.

Abstract

Data grows in value when joined and combined; likewise the power of voice grows in ensemble. With 15 UK choirs, we explore opportunities for bottom-up data governance of a jointly created Choral AI Dataset. Guided by a survey of chorister attitudes towards generative AI models trained using their data, we explore opportunities to create empowering governance structures that go beyond opt in and opt out. We test the development of novel mechanisms such as a Trusted Data Intermediary (TDI) to enable governance of the dataset amongst the choirs and AI developers. We hope our findings can contribute to growing efforts to advance collective data governance practices and shape a more creative, empowering future for arts communities in the generative AI ecosystem.

My Voice, Your Voice, Our Voice: Attitudes Towards Collective Governance of a Choral AI Dataset

TL;DR

The paper addresses the risk of artists losing control over their contributions when used to train AI models and proposes collective governance as a remedy. It introduces the Choral Data Trust Experiment, a case study where 15 UK choirs contribute to a Choral AI Dataset and explore governance via a Trusted Data Intermediary (TDI). Through surveys and Polis analyses, the study uncovers nuanced consent preferences, group-level credit expectations, and licensing patterns, informing the design of governance mechanisms. The work delivers concrete governance artifacts—an established TDI as a legal hub, a Performance Rights Agreement reflecting choir preferences, and a Data Rights Mandate for enforcing individual data rights—offering a scalable blueprint for empowering arts communities in the generative AI ecosystem.

Abstract

Data grows in value when joined and combined; likewise the power of voice grows in ensemble. With 15 UK choirs, we explore opportunities for bottom-up data governance of a jointly created Choral AI Dataset. Guided by a survey of chorister attitudes towards generative AI models trained using their data, we explore opportunities to create empowering governance structures that go beyond opt in and opt out. We test the development of novel mechanisms such as a Trusted Data Intermediary (TDI) to enable governance of the dataset amongst the choirs and AI developers. We hope our findings can contribute to growing efforts to advance collective data governance practices and shape a more creative, empowering future for arts communities in the generative AI ecosystem.

Paper Structure

This paper contains 8 sections, 8 figures, 1 table.

Table of Contents

  1. Introduction & Related Work
  2. The Choral Data Trust Experiment
  3. Building Capacity for Collective Data Governance
  4. Data Preferences Survey Findings
  5. Licence Preferences Polis Findings
  6. Prototyping Novel Governance Mechanisms
  7. Future Work
  8. Appendix / supplemental material

Figures (8)

  • Figure 1: Schematic (left) and image (right) depicting the recording setup for collection of the Choral AI Dataset, with a multi-microphone array capturing 8 close-range microphones for soloists, 4 room microphones and a first-order ambisonic microphone
  • Figure 2: Changes in comfort levels around the use of the Choral AI Dataset to train models by the exhibition artists (left) and other potential users (right)
  • Figure 3: Changes in preferences around crediting for individual contribution (left) and choir contribution (right) by future users of the Choral AI Dataset
  • Figure 4: Polis statements with the highest levels of disagreement among preference groups
  • Figure 5: Polis statements with the highest levels of agreement that informed recommendations for Choral AI Dataset licence terms and further investment in the Trusted Data Intermediary
  • ...and 3 more figures