OpenAI wins US backing in New York Times AI copyright fight

Washington stepping in on OpenAI’s side is a major moment in the battle over what it takes to build frontier models. The Justice Department is not just weighing in on one lawsuit, it is telling courts that using copyrighted text for training should usually count as fair use and that restricting it would damage US science, prosperity and national security.([news.bloomberglaw.com](https://news.bloomberglaw.com/us-law-week/trump-administration-backs-openai-in-ny-times-copyright-suit)) That gives large labs political cover to keep training on broad web-scale corpora instead of trying to stitch together thousands of fragmented licenses.

For the race to AGI, that stance effectively blesses the data strategy of the leading US labs. It lowers perceived legal risk around scaling training sets and makes it harder for content owners to force a pay-per-work training regime that only the richest companies could afford. That does not end the copyright fight, but it signals that at least one branch of the US government sees unconstrained training as a national priority. In the near term, this encourages continued investment in massive models and may dissuade cautious incumbents from slowing or geofencing their training pipelines while the case plays out.

The bigger question is whether Congress or other jurisdictions respond with new rules that claw back some control for rights holders. If they do not, this brief will look like a green light for very aggressive scaling of both data and compute.