TL;DR
The White House told Congress to let judges decide whether training AI on copyrighted work is theft. Most creators don't know it happened.
MSM Perspective
Nelson Mullins' legal analysis identified the copyright deferral as the framework's most consequential provision, noting it directly contradicts Sen. Blackburn's draft bill which would declare AI training on copyrighted works categorically outside fair use.
X Perspective
AI policy accounts are split: industry voices are relieved the framework avoided a legislative copyright crackdown, while creator-rights advocates like Ed Newton-Rex call it 'pathetic' and a deliberate dodge.
The White House's national AI legislative framework, released Thursday, covered seven policy areas. Yesterday's coverage focused on federal preemption of state AI laws. The provision that may matter more in five years got less attention: copyright.
The administration's position is that training AI models on copyrighted material is probably lawful. But rather than codify that view, the framework tells Congress to stay out and let courts resolve the fair use question. [1]
That is a choice disguised as a non-choice. By deferring to the judiciary, the White House avoids picking a legislative fight with the creative industries while giving AI companies exactly what they want: time. Every month the courts deliberate is another month of training on copyrighted works without clear legal consequence.
The framework also supports creating collective licensing frameworks so rights holders can negotiate with AI companies without triggering antitrust liability. [1] That sounds reasonable until you consider the power asymmetry. OpenAI, Google, and Anthropic have the legal teams and capital to structure licensing deals on their terms. Individual writers, illustrators, and musicians do not.
Ed Newton-Rex, the former Stability AI executive who resigned over copyright concerns, called the framework "pathetic" and "a non-starter." [2] Sen. Blackburn's competing TRUMP AMERICA AI Act takes the opposite position, declaring AI training on copyrighted works categorically outside fair use. [1] The two documents agree on almost everything else. On copyright, they are fundamentally opposed.
The silence from the creative community is the story within the story. Most working creators do not track White House policy frameworks. They learn about copyright erosion when their income changes, not when a four-page PDF appears on whitehouse.gov. By the time the courts rule, the training will be done, the models will be deployed, and the question will be less about permission than about compensation for a fait accompli.
The administration believes the courts will sort it out. The courts believe Congress should legislate. Congress is busy with a war. The AI companies are busy training.
-- DAVID CHEN, Beijing
Sources & X Posts
News Sources