Is it legal to train AI models on copyrighted books? It’s complicated
TL;DR
Most published authors have, without their knowledge or consent, contributed to the development of the same AI tools that threaten to undermine their livelihoods. That seems illegal, right? Most published authors have, without their knowledge or consent, contributed to the development of the same AI tools that threaten to undermine their livelihoods. That seems illegal, right?
Nauti's Take
The opportunity sits in the clarification itself: once courts define which training corpora need a licence, authors gain a basis for compensation and labs gain planning certainty. The risk falls on everyone training or fine-tuning on book data in the meantime, because provenance now weighs heavier in court than good intentions.
Teams building their own models should document every source, and teams buying tools should read the vendor's licensing statement closely.
Sources