All
Ai Content Generator
Ai Picture
Tell Your Story
World NewsSpawning Aims to Build More Ethical AI Training Datasets
Spawning Aims to Build More Ethical AI Training Datasets
Source.Plus Project Curates Non-Infringing Media for AI Model Training
The debate surrounding the ethics of training generative AI models continues to grow, with implications for artists and creators. Spawning AI, founded by Jordan Meyer and Mathew Dryhurst, has launched the Source.Plus project to address these concerns. The project aims to curate "non-infringing" media for AI model training, offering artists more control over how their works are used online. With a dataset of nearly 40 million public domain images and images under the Creative Commons' CC0 license, Source.Plus provides a high-quality alternative for developers while respecting data rights.
Rights Management:
Training generative AI models, particularly those focused on art generation, requires large quantities of relevant data. Some developers argue that fair use allows them to scrape data from public sources, regardless of copyright status, while others compensate or credit content owners. Spawning's CEO, Jordan Meyer, believes that there is no settled best approach yet. Source.Plus aims to provide artists and rights holders with control over their data and offer developers responsibly sourced alternatives.
"AI training frequently defaults to using the easiest available data — which hasn't always been the most fair or responsibly sourced. Artists and rights holders have had little control over how their data is used for AI training, and developers have not had high-quality alternatives that make it easy to respect data rights."
Source.Plus: A Universal "Opt-In" Platform:
Source.Plus is designed to be a universal "opt-in" platform, making it easy for rights holders to offer their media for AI training on their own terms. The initial dataset, PD/CC0, consists of high-quality images suitable for commercial or research applications. Spawning's existing tools for art provenance and usage rights management serve as the foundation for Source.Plus, ensuring that data rights are respected throughout the training process.
Ensuring Quality and Ethical Sourcing:
Spawning takes a proactive approach to ensure the quality and ethical sourcing of the Source.Plus dataset. The platform filters images for opt-outs and artist training preferences, providing provenance information about how and where the images were sourced. It excludes images that aren't licensed under CC0, including those requiring attribution. Spawning also monitors for copyright challenges from sources where creators may not be responsible for indicating the copyright status.
Compensation and Transparency:
Compensation for creators contributing to generative AI training datasets has been a contentious issue. Spawning aims to provide fair compensation and transparency for artists. Once Source.Plus expands beyond the beta phase, artists and rights holders will have the opportunity to set their own prices per download. Spawning charges a flat rate, ensuring artists receive the majority of the revenue. Additionally, a subscription plan called Source.Plus Curation allows users to manage collections of images privately and access premium features.
Spawning's Source.Plus project is a promising step towards building more ethical AI training datasets. By empowering artists and rights holders to control how their works are used in AI training, Source.Plus offers a viable alternative to platforms perceived as exploiting creators. With its commitment to responsible sourcing, compensation, and transparency, Spawning aims to create a more equitable generative AI ecosystem. As the project expands to include other types of media, the impact on artists and the AI community will be closely watched.
Related Posts
© 2024 Invastor. All Rights Reserved
User Comments
User Comments
There are no comments yet. Be the first to comment!