Warner Bros. Discovery Joins AI Copyright Fight Against Midjourney

Another major studio just jumped into the AI copyright wars. Warner Bros. Discovery is suing Midjourney, joining Disney and NBCUniversal in a growing legal coalition challenging the unauthorized use of studio-owned IP to train and power generative AI systems.

WBD’s complaint is not vague or speculative. The lawsuit, filed in U.S. District Court in Los Angeles, accuses Midjourney of “brazenly” distributing images and videos built off its iconic characters, including Batman, Superman, Wonder Woman, Scooby-Doo, and Rick and Morty. The studio calls this “mass theft,” alleging the company developed its commercial AI model by ingesting thousands of copyrighted works without consent and now offers that IP to paying users in an easily searchable, high-fidelity format.

This goes beyond questions about training data. WBD says the infringement is clearly visible in the outputs. The complaint includes examples where prompts like “Batman, screencap from The Dark Knight” return images that closely resemble Christian Bale’s version of the character. Another example shows a 3D Bugs Bunny image that mirrors his appearance in Space Jam: A New Legacy, with few, if any, changes.

The argument is that these outputs are not simply inspired by original work. They are competitive with officially licensed products like posters, prints, and other merchandise. According to WBD, this undermines the commercial value of their IP and diverts demand away from licensed goods.

Midjourney has previously argued that its platform is simply a tool that helps users express creative ideas, and that any copyrighted content used to train its AI models should fall under fair use. The company also pointed to DMCA procedures that allow rights holders to request takedowns of infringing content. But WBD is not targeting users or content removals. It is going after the business model itself.

The complaint highlights Midjourney’s four-tier subscription model, ranging from $10 to $120 per month. Warner argues that access to these infringing images is the primary driver of those subscriptions, which means the company is directly profiting from its IP without authorization.

While courts have yet to rule definitively on whether training AI models with copyrighted content qualifies as fair use, earlier cases offer some clues. In a case against Amazon-backed Anthropic, a judge noted that the training process may be protected under fair use, but left the door open for liability tied to the distribution of infringing outputs. Anthropic settled before a final ruling, but the legal framework that is taking shape could be useful to studios.

WBD is seeking either the profits Midjourney made from its allegedly infringing activity or statutory damages of up to $150,000 per infringed work. If successful, the outcome could result in enormous damages and potentially reshape how generative AI companies handle copyright.

So far, Paramount Skydance, Amazon MGM, Apple Studios, Sony Pictures, and Lionsgate have not joined the lawsuits. But many of these companies have made public moves in the AI space, and the outcome of this case could influence whether they jump in or stay quiet.

This fight is no longer theoretical. The biggest content owners in Hollywood are now aligned and actively challenging the foundational practices of one of AI’s leading platforms. Whether the courts see this as fair use or a commercial violation of copyright law will have major consequences across tech and media.