The article reports that a U.S. federal judge has allowed a class-action lawsuit brought by authors — including George R.R. Martin — against OpenAI to move forward. The lawsuit claims that OpenAI trained its large language models using copyrighted works without permission, and that its AI-generated content produced outputs substantially similar to those authors’ protected texts.
A key example cited in the case involved a prompt given to OpenAI’s system: the AI was asked to outline a sequel to Martin’s “A Clash of Kings” that diverged from his “A Storm of Swords”. The model proposed a new book titled “A Dance with Shadows”, complete with plot elements like “ancient dragon-related magic”, a Targaryen relative named Lady Elara seeking the Iron Throne, and a rogue sect of the Children of the Forest. The judge deemed these details sufficiently similar to Martin’s work to justify letting the case proceed.
The crux of the dispute revolves around whether using copyrighted works as training data (and generating derivative- or inspired-works) constitutes infringement, or whether it is protected under fair use. OpenAI holds that its use of publicly available text is lawful; the authors argue that the resulting AI output essentially replicates their creative world and characters without consent. The judge’s decision does not determine guilt, but allows the plaintiffs to pursue their claims in court.
In short, this development underscores the growing legal exposure for major AI developers: when AI output bears strong resemblance to specific copyrighted texts, the line between “inspired by” and “too close for comfort” is being tested in court. The outcome of this case could have wide implications for how training data is sourced and how AI-generated content is governed.
