Judges Appear Open to Undoing OpenAI Win in Copyright Suit (1)

Two media outlets seeking to revive digital copyright claims against OpenAI Inc. over its AI-training process met a receptive Second Circuit panel as the judges considered whether the publishers suffered the necessary harm to bring their lawsuit.

An “age-old” truth in copyright law “is reproduction itself violates the copyright—that’s part of what the right is,” Judge Richard C. Wesley said Wednesday.

A district judge in the fall of 2024 dismissed the publishers’ suit alleging OpenAI violated the Digital Millennium Copyright Act by removing author names and titles from their content before using it as artificial intelligence training material, finding they hadn’t suffered a concrete injury that would allow them to sue.

OpenAI’s counsel, Luke McCloud of Williams & Connolly LLP, pushed back, saying the district court was right to toss the lawsuit because the publishers lacked evidence their works were disseminated to create such an injury.

“You’re holding hard to, and your brief holds hard to, the lack of their evidence, that really, with all due respect to you, doesn’t impress me much,” Wesley responded.

“You haven’t had much discovery. I mean, you’re talking about facts behind a screen,” he said during the oral argument before the US Court of Appeals for the Second Circuit in New York.

Wesley continued to grill McCloud about whether OpenAI made copies of the works in the process of removing the publishers’ information. When McCloud began to respond that he didn’t know the technical details, Wesley interrupted him to say, “I would have thought you would, so that you could stand up there and say that there’s no proof whatsoever.”

Raw Story Media Inc. and AlterNet Media Inc. sued OpenAI in 2024. The lawsuit didn’t allege OpenAI infringed their copyrights, but instead accused the AI giant of violating the DMCA by intentionally removing copyright management information including author names, titles, copyright notices, and terms of use from their content.

Judge Colleen McMahon dismissed the case, ruling that removing of such information—absent dissemination of those works—doesn’t qualify as an adverse effect needed to establish standing to sue. She later denied the outlets’ bid to amend their complaint, saying their proposed changes didn’t make their alleged harm any more concrete.

The publishers argued on appeal they have standing because the harm is similar to that of copyright infringement. By removing CMI and creating CMI-less copies, OpenAI “harms creators’ incentives to make new works and thus undermines the very purpose of copyright law,” they said.

McCloud on Wednesday argued the publishers have a “traceability problem” regarding the DMCA because the removal of CMI didn’t cause the copying, the copying caused the removal of CMI.

Even if the violation isn’t traceable to the DMCA, Judge Dennis Jacobs told McCloud, “you were just granting to Judge Wesley that that alone could be a copyright infringement, and then you don’t need the CMI.”

McCloud contended the outlets can’t bring the claim Jacobs is alluding to because their case asserts a DMCA claim, not an infringement claim.

“If I tear the cover off of my copy of ‘Harry Potter,’ I have not infringed any property rights of J.K. Rowling,” he said. “Now, if that leads to eventual copyright infringement, there may be a claim both under the DMCA and under the copyright statutes.” This case is different because the publishers don’t have registered copyrights, McCloud said.

The outlets’ counsel, Stephen Stich Match of Loevy + Loevy, responded that the relevant question is whether they’ve suffered harm.

“If we were to register our copyrights tomorrow, our harm would be the same as it is today,” he said, “the making of a new copy and the violation of that right.”

Judge Eunice C. Lee also sat on the panel.

The case is In Re: OpenAI Inc. Copyright Infringement Litigation, 2d Cir., No. 25-1756, hearing held 3/18/26.