Judge dismisses news outlets' copyright lawsuit against OpenAI

Media outlets hoping to corner OpenAI in court just hit a major legal roadblock.

On Thursday, U.S. District Judge Colleen McMahon dismissed a copyright lawsuit brought by digital news publishers Raw Story and AlterNet against OpenAI. The core reason: the plaintiffs failed to show any concrete injury.

A clever legal sidestep falls flat

When Raw Story and AlterNet filed their complaint in February, they tried an unconventional legal tactic. Instead of suing OpenAI for outright copyright infringement, they accused the company of stripping copyright management information (CMI)—like author names and article titles—from thousands of articles before feeding them into training datasets.

They argued this practice violated the 1998 Digital Millennium Copyright Act (DMCA) and claimed an injunction was needed because ChatGPT might spit out their copyrighted text without proper credit.

Judge McMahon saw right through the maneuver. In her ruling, she pointed out that given the sheer scale of the training data, the likelihood of ChatGPT regurgitating plagiarized content from the plaintiffs' articles is remote. Factual information inside news stories, she noted, isn't protected by copyright anyway.

The judge also addressed past examples where ChatGPT reproduced text verbatim. She observed that those results cannot be replicated on current versions, treating verbatim output as an occasional bug rather than intended design.

McMahon captured the real dispute plainly: what plaintiffs truly seek redress for is not the exclusion of metadata, but the fact that OpenAI used their work to develop ChatGPT without paying for it.

While the judge dismissed the suit, she left the door open for the publications to file an amended complaint—provided they can prove it wouldn’t be an exercise in futility. The plaintiffs' attorney expressed confidence they could fix the issues, while OpenAI maintained that its models rely on publicly available data protected by fair use.

Why it matters

This ruling deals a significant blow to publishers trying to find technical workarounds to penalize AI developers.

For creators and media companies, taking the DMCA route without evidence of actual, real-world commercial harm is now a proven dead end. Courts are signaling that simply having your content ingested into a black-box training run does not automatically equal a compensable injury. You have to prove the model is directly copying you and cannibalizing your business.

For AI labs, McMahon’s perspective strongly bolsters the industry’s fair use argument. By recognizing that modern LLMs synthesize training data rather than copy it, the court reinforced the distinction between learning from the open web and pirating content.

That said, OpenAI isn't entirely off the hook. This case was dismissed largely because of how the plaintiffs framed their claim. Tougher battles remain, particularly The New York Times lawsuit, which attacks OpenAI head-on for allegedly creating a market competitor using millions of proprietary articles. But for now, the courtroom bar for stopping AI training has been set much higher.