US government sides with OpenAI on issue of training LLMs on copyrighted material
TechCrunch · 2026-09-02
The U.S. government has sided with OpenAI in the debate over using copyrighted material to train large language models (LLMs), filing a brief that defends the practice. The brief argues that the United States has a strong interest in developing a robust and competitive AI industry and that constraining LLM development due to a misunderstanding of fair use doctrine would hinder creative progress, American prosperity, and economic mobility. This intervention references an executive order signed by President Donald Trump and comes as many publishers, like The New York Times, contend that AI companies illegally train models on their copyrighted works without permission. While not a ruling, this stance by the Trump administration could carry weight in ongoing legal battles. Past cases have largely favored AI companies, focusing on issues like the use of illegal shadow libraries rather than the training itself.