US Court Evaluates Copyright Fair Use in AI Model Training
A federal judge is considering whether Anthropic’s use of copyrighted song lyrics to train its AI chatbot Claude constitutes fair use under US law.
In a case brought by major music publishers including Universal Music Group, Concord, and ABKCO, Anthropic argued that its use of song lyrics for AI training creates transformative content and does not diminish market demand. Publishers contend that the practice enables the generation of competing synthetic lyrics, threatening the livelihoods of songwriters.
US District Judge Eumi Lee heard arguments in San Jose but reserved judgment. The legal debate highlights the ongoing uncertainty surrounding AI model training, as previous rulings in the Northern District of California have reached conflicting conclusions regarding whether such data usage qualifies as fair use.
Internationally, India is exploring a different regulatory approach. The Department for Promotion of Industry and Internal Trade (DPIIT) has proposed a mandatory licensing framework, often referred to as "one nation, one licence," which would require AI developers to pay royalties for commercialized models. The legality of AI training remains a subject of pending litigation in India as well.