AI Copyright Training

Anthropic has asked a California federal judge to rule that its use of copyrighted materials to train Claude qualifies as transformative fair use, comparing the AI's training process to how humans learn by reading and absorbing themes. The filing stands apart from the $1.5 billion class-action settlement in Bartz v. Anthropic, where the claims deadline passed on March 30, 2026, and a fairness hearing is scheduled for May 14, 2026, in San Francisco federal court.

Core event/development: The Trump administration released the National Policy Framework for Artificial Intelligence on March 20, 2026, outlining seven legislative priorities: child protection, infrastructure/economic growth, intellectual property, content restrictions, innovation via regulatory sandboxes, workforce training, and federal preemption of state AI laws imposing "undue burdens."[1][5][7][15] Senator Marsha Blackburn introduced the Trump America AI Act, which includes mandatory bias audits for high-risk AI systems (e.g., in treatment recommendations, insurance), copyright-based training data rules, and transparency obligations, while preserving sectoral laws like FDA, HIPAA, and state health privacy.[1]