By Shilpi Saurav Sharan and Nitika Sinha
On June 11, 2025, [1]The Walt Disney Company and Universal Studios jointly filed a landmark federal lawsuit against Midjourney, an artificial intelligence company known for its sophisticated image-generation tools. The complaint, lodged in the U.S. District Court for the Central District of California, accuses Midjourney of “massive and deliberate copyright infringement” for allowing users to create highly realistic depictions of proprietary characters such as Elsa, Shrek, Darth Vader, and Homer Simpson[2]—without license or authorization.
The case is not merely a copyright dispute; it is a direct challenge to the core business model of generative AI companies that train on publicly scraped internet content. It represents a pivotal moment in the evolving clash between creative ownership and AI-driven innovation.
The Allegations: A “Bottomless Pit of Plagiarism”
At the heart of the complaint is the assertion that Midjourney trained its AI models using tens of thousands of copyrighted images belonging to the plaintiffs, sourced without permission. Disney and Universal argue that this resulted in outputs that are not only derivative but “virtually indistinguishable” from the originals—violating both the spirit and letter of U.S. copyright law.
According to the studios, cease-and-desist letters sent to Midjourney were ignored, and the company continued enabling users to prompt images of copyrighted characters and scenes. This, the plaintiffs claim, constitutes willful infringement, warranting enhanced statutory damages under 17 U.S.C. § 504(c).
The plaintiffs are seeking both injunctive relief—demanding Midjourney cease generating or training on their content—and the installation of proactive copyright filters to prevent future misuse.
Legal Context
Though this marks the first significant lawsuit from major Hollywood studios targeting a generative AI firm, it emerges within an increasingly active and evolving legal environment surrounding AI and intellectual property:
- Getty Images v. Stability AI (2023): Getty sued Stability AI in the U.K. and U.S. for using millions of watermarked photographs to train its model without a license. The court in the U.K. denied an early dismissal, signaling a willingness to hear the matter on its merits.
- The New York Times v. OpenAI & Microsoft (2023): The Times claimed its journalism was ingested and repurposed by large language models, resulting in outputs that replicated its copyrighted reporting nearly verbatim. The suit brought renewed focus on fair use and data licensing.
- Andersen v. Stability AI (2023): A group of artists filed a class-action suit arguing that their personal styles and images had been copied and reconstructed through generative tools. Though partially dismissed, the case introduced the concept of “style as property” into the copyright debate.
These cases—and now Disney v. Midjourney—highlight the unresolved legal question: Does training an AI on copyrighted material, even without copying the exact image, constitute infringement?
Fair Use or Overreach? The Stakes for AI Developers
Midjourney, like other AI companies, may invoke the fair use doctrine, arguing that its use of copyrighted works for training constitutes a transformative purpose, with no market substitution. However, courts have traditionally limited fair use when the resulting output undermines the original work’s commercial value—particularly in the entertainment sector, where licensing is a multi-billion-dollar industry.
What sets the Disney and Universal lawsuit apart is the clear identifiability of the characters in question. The filing highlights direct visual comparisons, showing that the AI-generated images closely replicate the originals with minimal variation. This diminishes the likelihood that a court will see such use as transformative or de minimis.
A Turning Point for Generative AI Regulation
If the studios succeed, the implications could be far-reaching:
- Training Data Scrutiny: AI companies may be legally compelled to disclose training sources and obtain licenses for all copyrighted content—similar to how streaming platforms must license music, scripts, and media assets.
- Mandatory Safeguards: Courts could require the implementation of technical safeguards—akin to YouTube’s Content ID system—to prevent real-time generation of infringing material.
- Industry-wide Licensing Regimes: The lawsuit could accelerate the development of AI-specific licensing frameworks, where IP holders are compensated for inclusion in model training or prompt-based use.
On the other hand, a ruling in Midjourney’s favor could embolden AI firms to train on publicly available content with fewer constraints, dramatically shifting the balance of power between rights holders and developers.
A Precedent in the Making
The Disney and Universal lawsuit against Midjourney mark the most aggressive and consequential legal action to date in the realm of AI-generated content. It is a case that squarely confronts the core legal tensions surrounding generative AI: innovation versus infringement, fair use versus appropriation, and automation versus authorship.
As the courts begin to draw lines in this uncharted territory, the outcome of this case will almost certainly influence future legislation, model development practices, and the broader relationship between artificial intelligence and creative ownership.
For stakeholders across industries, whether entertainment, publishing, law, or technology, this is a case that merits close and continuous attention.
Feba Sara Vinu, Former Junior Associate Advocate at S.S.Rana & Co. has assisted in the research of this article.
[1] https://media.npr.org/assets/artslife/movies/misc/midjourney.pdf