Toronto AI Startup Faces Legal Headwinds as U.S. Court Rejects Motion to Dismiss Copyright Suit

  s-legal-headwinds-as-u-s-court-rejects-motion-to-dismiss-copyright-suit.html
  Media and Entertainment on Fri, Nov 14th, 2025 by Toronto Star

Toronto AI Startup Faces Legal Headwinds: U.S. Court Rejects Motion to Dismiss Copyright Suit

A Toronto‑based artificial‑intelligence company has found itself on the wrong side of the U.S. courts, as a federal judge in the Southern District of New York refused to dismiss a copyright infringement lawsuit filed by a U.S. publisher. The decision, reported by The Star, marks a significant moment in the ongoing legal battle over how AI‑generated text intersects with traditional copyright law.

The Parties Involved

The plaintiff, The Global Media Group (GGM), is a U.S. publisher with a portfolio that includes the New York Herald and Time‑line Magazine. GGM alleges that the Toronto startup, Cortex AI, used large swaths of its copyrighted content to train its language‑modeling system, which it offers as a white‑label solution to other businesses.

Cortex AI, founded in 2020 by a team of former machine‑learning researchers from both Toronto and Toronto’s neighboring university, markets itself as a “next‑generation content‑generation platform.” Its flagship product, CortexWriter, is marketed to marketing agencies, academic institutions, and even legal firms. The company claims that it only uses publicly available text and that its system performs a transformative use that falls under fair‑use protections.

The Legal Groundwork

GGM’s suit, filed on March 2, 2025, alleges that Cortex AI used a proprietary dataset comprising 3.2 million pages of copyrighted articles without permission. The complaint claims that the startup copied, transformed, and redistributed large portions of these texts in the creation of its AI models. The plaintiff is seeking $50 million in damages, citing lost revenue and brand dilution.

In response, Cortex AI filed a motion to dismiss on April 10, arguing that the claim is barred by the copyright exhaustion doctrine and that the alleged use is transformative enough to qualify for fair‑use. The motion also cited a recent California appellate decision that recognized certain “in‑context” uses of copyrighted material as transformative, suggesting that the company’s data‑collection process does not infringe.

The Court’s Decision

Judge Laura M. Hayes denied the motion on April 25. In a detailed memorandum, she noted that while fair‑use is a complex defense, the facts presented by GGM indicate a strong likelihood of infringement. Hayes wrote that the plaintiff had clearly documented that the text was extracted from its digital archives, and that Cortex AI’s training process involved feeding the text into its neural network in a way that “reproduced the original works in a substantial manner.”

She also noted that the transformative nature of the use is “unclear and speculative.” The startup’s argument that the data is “public domain” is insufficient, Hayes explained, because GGM’s content is not in the public domain and the data was extracted from the publisher’s proprietary subscription services.

The judge’s ruling does not preclude a fair‑use defense at trial, but it establishes that the case will proceed to the discovery phase. “The parties have a real chance to present their evidence and argue the merits,” Hayes said.

Implications for the AI Community

The decision is part of a growing list of legal challenges that AI firms face as they continue to use large corpora of text for model training. Several high‑profile lawsuits, including those involving OpenAI and Google, have centered on similar issues of data sourcing and copyright infringement.

Industry experts warn that the outcome of this case could set a precedent for how courts view the transformation of copyrighted text in AI models. If GGM wins, it could reinforce the requirement that AI companies obtain licenses for the content they use, potentially increasing costs for startups. Conversely, if Cortex AI successfully argues a fair‑use defense at trial, it could provide a roadmap for other AI firms looking to avoid legal entanglements.

What Happens Next?

Cortex AI has announced that it will continue to defend itself and is already engaging a U.S. litigation team. The startup also plans to file a request for a declaratory judgment that its data‑collection methods do not infringe copyright, a move that has been used in prior cases to clarify legal standing.

The judge’s ruling also opens the door for a potential summary judgment on the fair‑use claim, depending on how the evidence is presented in discovery. Both parties have been active on social media, with GGM tweeting that “the future of AI must respect the rights of creators.” In response, Cortex AI’s CEO, Mira Patel, posted a statement asserting the company’s commitment to ethical data use and collaboration with content creators.

Additional Context

The Star article links to a PDF of the court’s memorandum, which offers a deeper dive into the legal reasoning. A related piece, cited in the article, examines a similar lawsuit involving a U.S. news aggregator, providing a broader view of the litigation landscape. Furthermore, the article references an interview with a copyright law scholar, Professor Elena Ramirez of Columbia Law School, who weighed in on the transformative use question.

As the legal battle unfolds, both the AI sector and the publishing industry will be watching closely. The outcome could reshape how AI developers approach data sourcing, potentially prompting new industry standards and licensing models.

In the meantime, the Toronto startup’s fate will hinge on its ability to navigate the complex interplay between innovation, data, and intellectual property law—an endeavor that will test not only its legal strategy but also its core mission to democratize AI technology for businesses worldwide.