The legal battle between the generative artificial intelligence startup Midjourney and a coalition of major Hollywood studios has entered a contentious new phase as the AI company seeks to compel the entertainment giants to disclose their own internal use of AI technology. This strategic maneuver by Midjourney is designed to bolster its "fair use" defense by attempting to demonstrate that the very entities suing for copyright infringement are utilizing similar technologies and methodologies behind closed doors. The dispute, which involves industry titans Disney, Universal, and Warner Bros., represents a pivotal moment in the intersection of intellectual property law and the rapidly evolving field of generative artificial intelligence.
The Genesis of the Litigation
The legal conflict began in June 2024 when Disney and Universal filed a joint lawsuit against Midjourney. The complaint alleged that Midjourney’s image-generation models were capable of producing near-perfect replicas of iconic, copyrighted characters without authorization. Evidence presented in the initial filings included AI-generated images of Bart Simpson, Darth Vader, and various Marvel superheroes—properties that represent billions of dollars in intellectual property value for the studios.
The litigation expanded in September 2024 when Warner Bros. Discovery filed a separate but similar lawsuit. This action focused on the unauthorized generation of images featuring DC Comics characters, such as Superman and Batman, as well as characters from the Harry Potter franchise. The studios collectively argue that Midjourney’s business model is built upon the mass scraping of copyrighted works, which are then used to train models that compete directly with the studios’ own commercial interests.
Midjourney has consistently maintained that its processes fall under the protection of the "fair use" doctrine. The startup argues that its AI models do not store copies of the images they are trained on but rather learn the underlying patterns and mathematical relationships between visual concepts. This process, according to Midjourney, is transformative and serves a different purpose than the original works.
The Discovery Dispute: Internal vs. Consumer-Facing AI
As the cases progressed into the discovery phase—the period where both sides must exchange relevant documents and evidence—a significant disagreement arose regarding the scope of information the studios must provide. Midjourney’s legal team has argued that to mount an effective defense, they require access to documents detailing how Disney, Universal, and Warner Bros. use generative AI in their own production pipelines.
A preliminary ruling by the presiding judge initially limited the scope of this discovery. The court ruled that the studios would only be required to provide information regarding generative AI usage that resulted in "consumer-facing" content, such as final theatrical releases, television broadcasts, or marketing materials. The court’s rationale was that internal experimentation or development tools were less relevant to the claims of market harm.
However, in a recent filing, Midjourney has moved to overturn this limitation. The startup contends that the current restriction is "unfair" and allows the studios to "cherry-pick" only the documents that support their narrative of market harm while suppressing evidence that could exonerate Midjourney. The core of Midjourney’s argument is that the studios are likely using AI for internal tasks such as storyboarding, concept art, and pre-visualization—processes that often involve training models on existing copyrighted materials.
The "Industry Custom" Defense
Midjourney’s legal strategy hinges on the concept of "industry custom." The startup posits that if the studios themselves are using AI models trained on unlicensed content for internal "ideation" or "storyboarding," then such practices should be considered a standard and acceptable use within the entertainment industry.
"The documents [the studios] are withholding are precisely those that would reveal whether, behind closed doors, they are doing exactly what they are suing Midjourney for doing," the startup stated in its latest filing. By revealing these internal practices, Midjourney hopes to demonstrate that the training of AI on copyrighted data is a necessary and transformative step in modern content creation, rather than a malicious act of infringement.
Furthermore, Midjourney is demanding that the studios reveal the full history of their employees’ interactions with the Midjourney platform. This includes not just the prompts that produced the allegedly infringing images of Darth Vader or Superman, but every prompt and resulting output generated by studio accounts. Midjourney suggests that this data will show the studios were using the tool for legitimate creative purposes, further complicating their claims of irreparable harm.
Official Responses and the "Fishing Expedition" Counter-Argument
The studios have pushed back forcefully against Midjourney’s demands. David Singer, the lead attorney representing the studios, has characterized Midjourney’s request for internal documents as a "fishing expedition." In legal terms, a fishing expedition refers to a broad, unfocused request for information in the hopes of finding something incriminating, rather than seeking specific evidence related to the case at hand.
Singer has clarified the studios’ position, stating that the litigation is not an attempt to stifle innovation or destroy the AI industry. "The studios do not seek to stop AI technology or even shut down Midjourney’s business," Singer remarked during a recent hearing. Instead, he emphasized that the goal is to protect the integrity of their intellectual property. The studios want Midjourney to "stop copying their movies and TV shows and to stop distributing, publicly displaying, publicly performing, and creating derivative works" without a licensing agreement in place.
The studios argue that their internal use of AI—if it exists in the manner Midjourney suggests—is fundamentally different because it occurs within a closed ecosystem where they own the underlying rights to the characters being used. They contend that there is a vast legal difference between a studio using its own IP to train an internal tool and a third-party startup scraping the entire internet to build a commercial product.
Supporting Data and Context: The Broader AI Legal Landscape
The conflict between Midjourney and Hollywood is not an isolated incident but part of a global wave of litigation targeting generative AI companies. According to data from legal analytics firms, there are currently over 30 high-profile copyright lawsuits pending against AI developers in the United States alone. These include:
- The New York Times vs. OpenAI and Microsoft: A landmark case alleging that millions of articles were used to train ChatGPT without permission.
- Getty Images vs. Stability AI: A suit focused on the unauthorized use of millions of stock photos, often resulting in AI images that still bear the Getty watermark.
- Artists vs. Midjourney and DeviantArt: A class-action lawsuit brought by visual artists who claim their unique styles are being "stolen" by AI models.
The outcome of the Midjourney-Hollywood dispute could set a significant precedent for how the "market harm" factor of fair use is evaluated. If Midjourney successfully proves that studios are using AI in similar ways, it may become harder for the plaintiffs to argue that Midjourney’s existence poses a unique existential threat to their business models.
Timeline of Key Events
- June 2024: Disney and Universal file the initial copyright infringement lawsuit against Midjourney in the Central District of California.
- September 2024: Warner Bros. Discovery files a separate lawsuit, which is later consolidated for discovery purposes with the Disney/Universal case.
- Late 2024: Midjourney files its initial response, citing fair use and the transformative nature of AI training.
- January 2025: The court sets the initial parameters for discovery, limiting studio disclosures to "consumer-facing" AI applications.
- June 2025: Midjourney files a motion to compel, seeking to expand discovery to include internal studio AI tools and all prompt history.
- Present: The court is expected to rule on the motion to compel in the coming weeks, a decision that will dictate the depth of transparency required from Hollywood’s major players.
Implications for the Creative Industries
The implications of this case extend far beyond the courtroom. If Midjourney prevails in expanding the discovery scope, it could force a level of transparency in Hollywood that many studios would find uncomfortable. The "behind-the-scenes" use of AI is a sensitive topic in the industry, particularly following the 2023 writers’ and actors’ strikes, where the use of AI was a central point of contention in labor negotiations.
If it is revealed that studios are heavily reliant on AI trained on unlicensed data for their internal creative processes, it could damage their leverage in future labor disputes and complicate their public-facing stance on copyright protection. Conversely, if the studios successfully block these discovery requests, it will reinforce the legal shield protecting internal corporate R&D from being used against them in infringement cases.
Furthermore, a victory for Midjourney on the "industry custom" argument could lead to a broader "de facto" legalization of AI training on copyrighted data, provided the output is sufficiently transformative. This would represent a seismic shift in intellectual property law, potentially moving the needle away from strict copyright enforcement and toward a more permissive environment for machine learning.
Conclusion and Current Status
As the legal teams for Midjourney and the Hollywood studios prepare for their next round of arguments, the tech and entertainment worlds remain closely attuned to the proceedings. The core question—whether a company can be held liable for "copying" when the output is a new, synthesized image—remains the most significant legal challenge of the 2020s.
The upcoming ruling on the discovery motion will be a bellwether for the case. Should the judge grant Midjourney access to the studios’ internal AI practices, the litigation could move toward a settlement as studios seek to avoid exposing their proprietary workflows. If the motion is denied, Midjourney will face an uphill battle in proving that its training methods are a standard industry practice. For now, the case stands as a high-stakes game of legal chess, with the future of creative IP and artificial intelligence hanging in the balance.
