Is it legal to train AI models on copyrighted books? It’s complicated

Quick take

Most published authors have had their copyrighted books used to train AI without their consent. This legal gray area centers on whether training models on copyrighted content constitutes fair use or an infringement.

Why it matters

Using copyrighted books to train AI tools directly challenges established copyright rules. For authors, this means their work is fueling AI systems that could compete with or replace their income streams, without compensation or approval. For AI builders and businesses, it raises legal risks and uncertainty over liability and the need to secure usage rights. For investors and operators, it signals potential costly litigation and regulatory challenges ahead. This tension forces a reckoning on how intellectual property law adapts to AI training practices, affecting who controls content and who profits from AI outputs.

AI Quick Briefs Editorial Desk