U.S. Government Backs AI Companies' Right to Use Copyrighted Data

In a Department of Justice filing, the U.S. government sided with AI developers on fair use grounds, arguing that training large language models on copyrighted material without permission falls within legal bounds. This position contradicts entertainment and publishing industries seeking licensing fees. The stance resolves a core friction point for AI labs: they can scale without negotiating thousands of licensing deals. It also crystallizes the political economy of AI—computational advantage flows to companies that can absorb legal risk while traditional IP holders face systematic devaluation. The filing converts a technical problem into a settled regulatory posture, at least temporarily.