Ziff Davis and CNET file second amended complaint against OpenAI for copyright infringement
Bot Mutiny |
A group of major digital publishers has filed a new complaint in New York federal court alleging OpenAI intentionally stripped copyright management information to train its language models.
Ziff Davis and a coalition of digital media brands filed a second amended complaint against OpenAI on August 17, 2026, in the Southern District of New York. The filing, documented as entry 1675 in the case In Re: OpenAI, Inc. Copyright Infringement Litigation, marks a significant escalation in the legal fight over how large language models ingest proprietary journalism. The document was first reported by CourtListener. The plaintiffs include Ziff Davis, Inc., IGN Entertainment, Everyday Health Media, Mashable, and CNET Media. They allege that OpenAI has built its business by relentlessly reproducing exact copies and creating derivatives of their works without authorization. Ziff Davis claims it produces nearly 2 million new articles and updates every year, including 5,000 product reviews. The complaint argues that OpenAI is using this specific, high-value content to power its Generative Pretrained Transformer (GPT) series. ## Systematic extraction and DMCA violations
The lawsuit isn't just about training data. It's about how OpenAI gets it. The complaint alleges that OpenAI knowingly stripped out copyright management information (CMI) from the articles it scraped. By removing this metadata, OpenAI allegedly distributed copies of the works while hiding their origin. The filing details a specific pattern of behavior. OpenAI allegedly copied text from Ziff Davis websites while violating and circumventing explicit written terms and technological controls. The publishers claim OpenAI then reproduced this content verbatim or in close paraphrase for third-party users. This directly harms the publishers' ability to monetize their work through advertising and product sales commissions. ## OpenAI rebuffed licensing talks
According to the complaint, the publishers tried to resolve this outside of a courtroom. Ziff Davis sent letters to OpenAI on February 5 and May 20, 2024, putting the company on notice. On August 23, 2024, Ziff Davis wrote to OpenAI again, providing evidence of infringement and requesting a meeting to discuss a licensing deal. OpenAI rebuffed the invitation to meet. The publishers argue that while OpenAI refuses to pay them, the tech company is simultaneously creating a market to license content from other publishers. This creates a two-tiered system where some data is paid for while other data is taken by force. ## The "move fast" defense
The complaint takes aim at OpenAI's corporate structure. It names seven different entities, including OpenAI, Inc., OpenAI GP, LLC, and OAI Corporation. The plaintiffs allege that OpenAI has abandoned its founding principle of openness. By ending its transparent publication practices, the company has concealed the extent of its misconduct. The lawsuit states that OpenAI is operating on the assumption that federal courts won't be able to redress these concerns before the tech becomes too entrenched. Ziff Davis claims that whole portions of OpenAI's models are themselves infringing copies of their work. The plaintiffs are demanding a jury trial. They're seeking relief for copyright infringement, violations of the Digital Millennium Copyright Act, and unjust enrichment under Delaware state law. Case 1:25-md-03143-SHS-OTW is now the record of a digital media industry fighting for its life against a company that allegedly sees journalism as free fuel for its software. The full docket and the primary document can be accessed via CourtListener.