New Nonprofit 'Fairly Trained' Certifies AI Companies Using Consented Data

Amidst copyright controversies, the organization aims to incentivize ethical data practices in AI training.

Ed Newton-Rex, former executive at Stability AI, has launched a nonprofit called 'Fairly Trained' to certify AI companies that train their models only on data whose creators have consented.
The move comes amid a growing debate over the use of copyrighted work to train AI systems, with several lawsuits filed against major AI companies like OpenAI and Meta.
Fairly Trained does not ask companies seeking certification to share their datasets for auditing, but instead requires written submissions detailing their data sources and due diligence processes.
Nine models had been certified by Fairly Trained at its launch, many of them made by AI companies in the music-generation space.
Fairly Trained charges fees for its certification service on a sliding scale based on the companies' annual revenue.