Empowering Artists in the AI Age
A project aims to give artists control over their creative contributions to AI.
Jennifer Ding, Eva Jäger, Victoria Ivanova, Mercedes Bunz
― 5 min read
Table of Contents
In the world of artificial intelligence (AI) and creativity, many artists worry about losing control over their work. When artists’ music or voices become part of AI training data, they often feel uneasy. They fear that their unique expression might be used without their permission or knowledge. While some solutions, like allowing artists to opt in or out, have been proposed, there's a growing interest in finding better ways to share power and decision-making.
The Choral Data Trust Experiment
To address this issue, a unique project, known as the Choral Data Trust Experiment, was initiated. The project brought together 15 choirs from across the UK to create a special dataset that AI models could learn from. This dataset was built through performances of a songbook composed by well-known artists Holly Herndon and Mat Dryhurst. To gather high-quality recordings, the project used advanced audio technology, which allowed for a better listening experience.
During the project, artists worked closely with researchers and engineers to create AI models that could understand and replicate choral music. However, the challenge was not just about the music and technical aspects; it also involved figuring out how to manage the data gathering and decision-making processes in a way that respected the choirs and their members.
The Role of Trusted Data Intermediaries
A key part of this project was the introduction of a Trusted Data Intermediary (TDI). This team was made up of art curators, legal experts, and a data steward. Think of the TDI as a referee in a game, ensuring that everyone plays by the rules and feels valued. They communicated with choir members about how their data might be used, giving them a say in the matter.
To engage with the choirs, the data steward hosted discussions and surveys, seeking to understand how choir members felt about their contributions and the governance of the dataset. These conversations were essential for capturing the thoughts and feelings of the choir members, which ultimately guided the design of the dataset's governance.
The Importance of Listening to Choir Members
One of the interesting discoveries from the discussions was how choir members felt about the term “data.” Many expressed discomfort with calling their artistic expressions “data,” as it felt impersonal and removed from the heart of what they do. Instead of viewing their contributions as mere numbers, they wanted recognition for their artistry.
Despite some initial resistance, it was clear that transparency played a crucial role in building trust. When choir members knew how their work would be used, they felt more comfortable. Interestingly, almost all choir members preferred to be recognized as part of a group rather than as individuals. Around 92% wanted their choir credited, showing that collective recognition mattered more to them than personal accolades.
Preferences for Data Sharing
When it came to sharing the Choral AI Dataset, opinions varied across the choirs. Some people were open to sharing their contributions widely, while others wanted to stay cautious and limit how their music and voices could be used. To figure this out, the project team set up a system for choir members to express their preferences, allowing them to vote on various scenarios regarding the use of their data.
Most choir members agreed that the dataset should be shared, but they wanted clear rules on how it could be used. They were particularly interested in ensuring that any use of their contributions for commercial purposes would not happen without their agreement. The feedback suggested that if the dataset were to be made publicly available, a non-commercial license would be the best way to move forward.
Creating Clear Guidelines and Legal Agreements
Based on the preferences expressed by choir members, the project aimed to create new governance mechanisms to ensure that everyone’s voice was heard. This included developing formal agreements that outlined how the data could be used and what rights choir members had concerning their contributions. By doing so, the TDI would act as a guardian, safeguarding the interests of the choir members.
The project also created a Performance Rights Agreement, which allowed choir members to set clear terms about how their data could be used. This agreement aimed to empower choir members by giving them a say in the ongoing management of the dataset and any future developments.
Future Directions and Ongoing Projects
The Choral Data Trust Experiment is still in progress, and many questions remain. How can collective governance be effectively managed when there are many contributors? What tools can help build trust and transparency among diverse groups? As the project continues, it hopes to explore these questions further, inviting other arts and AI communities to join the conversation.
The experience gained from the project will be invaluable for future efforts to empower artists and ensure that they are not just treated as mere data points in AI training. By finding ways to give artists a voice in how their work is used, the project hopes to create a better balance between creativity and technology.
Conclusion: Balancing Art and AI
In conclusion, the Choral Data Trust Experiment illustrates the importance of collective governance in the age of AI. By involving artists in the decision-making process, this project aims to foster an environment where creativity can thrive without sacrificing control. Through open conversations and transparent practices, the initiative shows that it's possible to create a space where artists and AI can coexist and collaborate.
As we move forward, understanding and addressing the concerns of artists will be essential. The project emphasizes that while technology continues to evolve, the human element—creativity, collaboration, and compassion—should remain at the forefront. So, whether you’re a choir member or just a fan of good music, the idea is to ensure that everyone’s voice counts in the growing world of generative AI. After all, who wouldn’t want a say in the chorus of their own life?
Original Source
Title: My Voice, Your Voice, Our Voice: Attitudes Towards Collective Governance of a Choral AI Dataset
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.
Authors: Jennifer Ding, Eva Jäger, Victoria Ivanova, Mercedes Bunz
Last Update: 2024-12-02 00:00:00
Language: English
Source URL: https://arxiv.org/abs/2412.01433
Source PDF: https://arxiv.org/pdf/2412.01433
Licence: https://creativecommons.org/licenses/by/4.0/
Changes: This summary was created with assistance from AI and may have inaccuracies. For accurate information, please refer to the original source documents linked here.
Thank you to arxiv for use of its open access interoperability.
Reference Links
- https://spawning.ai/have-i-been-trained#content
- https://huggingface.co/spaces/bigcode/in-the-stack
- https://www.ircam.fr/#
- https://www.serpentinegalleries.org/whats-on/holly-herndon-mat-dryhurst-the-call/
- https://holly.plus/
- https://proud-paprika-325.notion.site/Data-Card-13debad528fe809c94b0f463d14152b5
- https://datatrusts.uk/
- https://drive.google.com/drive/folders/1YWz-sJ-ZKqowPxuNObQcHLQzV0T9AJ49?usp=drive_link
- https://docs.google.com/document/d/1eYALwMy8zTtloRtWyJE4_Q4KVqYtj55e_BohlTsnYh0/edit?usp=sharing
- https://pol.is/report/r892azrash5nkrkscejbr
- https://creativecommons.org/licences/by-nc-sa/4.0/deed.en
- https://nips.cc/public/guides/CodeSubmissionPolicy
- https://neurips.cc/public/EthicsGuidelines