Ed Newton-Rex is the founder of Fairly Trained, a nonprofit certification organization for AI companies that use only licensed or consented training data, and a figure in the AI copyright debate. He previously served as VP of Audio at Stability AI.
Background and roles
Newton-Rex resigned as VP of Audio at Stability AI in November 2023 over disagreements about the company's position that training on copyrighted data without permission constitutes fair use. He subsequently founded Fairly Trained, a nonprofit that certifies AI companies whose models are trained only on licensed or consented data. The certification program offers a market-based mechanism for distinguishing models trained on licensed data.
Positions and statements
Newton-Rex argues that AI developers should obtain consent from or compensate creators whose work is used for training. His resignation from Stability AI marked a high-profile moment in the AI Copyright discourse. His advocacy intersects with ongoing litigation, including NYT v. Microsoft, OpenAI et al., and with broader policy questions about whether existing copyright frameworks are adequate for AI training, aligning with plaintiff-side arguments on training-data rights.
Relationships
- leads: Fairly Trained — founder of training-data certification organization
- related: AI Copyright — central advocate in the AI copyright debate
- related: NYT v. Microsoft, OpenAI et al. — his advocacy aligns with plaintiff-side arguments on training-data rights