A US judge ruled that Anthropic’s use of copyrighted books to train
AI models without author consent qualifies as “fair use,” marking the first court endorsement of such practices by a major tech firm. The decision sets a precedent in ongoing legal battles over AI training data.