Spawning wants to build more ethical AI training datasets | TechCrunch

Spawning AI has launched Source.Plus, a curated platform designed to offer artists greater control and fair compensation for their work in generative AI training. By focusing on a high-quality dataset of public domain and CC0-licensed images, the project aims to provide a reliable alternative to the often unregulated scraping methods used by major AI developers. This initiative addresses the critical need for ethical data sourcing, allowing creators to opt-in on their own terms while ensuring developers can access usable materials without legal ambiguity. The platform distinguishes itself through rigorous validation processes, excluding content with questionable licenses and employing classifiers and moderators to filter out harmful or non-consensual material. Unlike existing models that often rely on opaque metrics or minimal payouts, Source.Plus empowers rights holders to set their own prices for data usage. Spawning charges only a nominal flat fee, ensuring that the vast majority of revenue goes directly to the contributors, thereby establishing a more transparent and equitable financial framework for creative professionals. This approach is highly relevant to open_data as it challenges the prevailing narrative that open data must be free and unrestricted to be useful. Source.Plus demonstrates that open, high-integrity datasets can coexist with robust rights management and artist compensation. It offers a potential model for sustainable open ecosystems where data availability does not come at the cost of creator exploitation, providing a viable path for integrating ethical standards into the broader open data and AI development community.

Source: techcrunch.com
Published on 2024-06-12