Another publisher has joined the fast-growing line of plaintiffs testing how copyright law applies to generative AI. U.S. News World Report has sued OpenAI in the Southern District of New York, alleging the company used its content without authorization to train AI models and generate outputs that compete with or diminish the value of the publisher’s work. The case, U.S. News World Report, L.P. v. OpenAI, Inc. et al, adds another closely watched dispute to a litigation trend that is rapidly becoming one of the most consequential battles in technology and media law.

At a high level, these cases raise a core question: when AI developers ingest large volumes of copyrighted material to train models, does that qualify as lawful fair use, or does it require permission and compensation? Publishers bringing these suits generally argue that model training and AI-generated summaries or reproductions exploit protected expression and threaten traffic, subscriptions, and licensing markets.

The U.S. Supreme Court has wiped away a Fifth Circuit ruling that upheld a copyright verdict against Grande Communications Networks, sending the case back for reconsideration in light of the Court’s recent decision narrowing when internet service providers can be held liable for subscribers’ piracy. The move does not end the dispute, but it is an important reset in one of the closely watched lines of cases testing secondary copyright liability against broadband providers.

In practical terms, the justices granted, vacated, and remanded the case, directing the Fifth Circuit to take another look under a new liability framework.