Simplifying AI Training Data Management for Artists

Sydney Times

— 1 min read
Simplifying AI Training Data Management for Artists

Introduction

Spawning AI, founded by Jordan Meyer and Mathew Dryhurst, aims to empower artists to have more control over how their works are used online. Their latest project, Source.Plus, is designed to curate high-quality media for AI model training, starting with a dataset of nearly 40 million public domain and Creative Commons images.

Rights Management

  • The ethics of training generative AI models like Stable Diffusion and DALL-E 3 are under scrutiny.
  • Spawning's CEO, Meyer, believes that current approaches to data rights management in AI training are still evolving.
  • Source.Plus, in beta, is Spawning's platform to support art provenance and usage rights management.

Data Quality

  • Source.Plus offers a curated dataset of non-infringing CC0 images for commercial and research use.
  • By filtering out questionable licenses and monitoring for copyright issues, Source.Plus maintains a high-quality dataset.
  • Image classifiers are used to detect and filter out inappropriate content in the dataset.

Compensation

  • Compensation for artists contributing to AI training data has been a contentious issue.
  • Source.Plus proposes a flat rate fee for access to the dataset, with artists setting their own prices per download.
  • An additional subscription plan, Source.Plus Curation, offers advanced features for managing image collections.

Future Expansion

  • Spawning plans to expand Source.Plus beyond images to include audio and video datasets.
  • Discussions are ongoing with potential partners to make diverse data available on Source.Plus.
  • Spawning may develop its own generative AI models using data from Source.Plus.

Conclusion

  • Source.Plus offers a promising opportunity for artists to participate in the generative AI economy.
  • The platform aims to provide fair compensation to rights holders while respecting data rights.
  • Source.Plus addresses the growing demand for ethical and transparent AI training data management in the creative community.

In the rapidly evolving landscape of AI training data management, Source.Plus stands out as a proactive solution that prioritizes the interests of artists and content creators.

Read more