Warner Bros. Discovery has joined the growing pile-on of Hollywood studios taking generative-AI companies to court, filing a federal lawsuit that accuses Midjourney of letting its image and video tools reproduce the studio’s most valuable characters — from Superman and Batman to Bugs Bunny and Scooby-Doo — “as if it were its own.” The complaint, lodged this week in Los Angeles, paints a picture of an AI service that not only produces near-identical recreations of copyrighted characters on command, but does so at a scale the studio calls breathtaking.

The headline claim — and the examples

At its core, Warner Bros. Discovery (WBD) says Midjourney trained its models on copyrighted material without permission and then removed or weakened guardrails that previously blocked the easy reproduction of those works. The complaint includes a battery of side-by-side images and screenshots — classic storefront evidence in the new AI copyright fights — alleging that Midjourney’s outputs are close enough to the originals that ordinary consumers would readily recognize the studio’s characters. WBD also says Midjourney will generate those characters even when prompts don’t explicitly name them; a generic “classic comic book superhero battle,” the suit alleges, produced recognizable images of Superman, Batman and Flash.

Warner Bros. Discovery accuses Midjourney of reproducing, displaying and distributing “unauthorized derivatives,” and uses striking language in the filing: “It is hard to imagine copyright infringement that is any more willful than what Midjourney is doing here,” the suit declares, arguing the company exploited WB’s IP to attract subscribers and profit. The studio is seeking damages, disgorgement of profits, and an injunction that would bar Midjourney from creating, displaying or distributing works derived from Warner Bros. properties — or from offering tools without “meaningful copyright protection measures.”

Hollywood’s broader legal offensive

This lawsuit is not happening in a vacuum. Earlier this summer, Disney and NBCUniversal (Universal) filed a similar complaint accusing Midjourney of operating a “virtual vending machine” that spits out “endless unauthorized copies” of their copyrighted creations. That pair of suits signaled a new, coordinated push from legacy media companies to hold AI labs accountable for the datasets and behavior of their models. Warner’s filing puts another major studio behind that effort and raises the stakes for generative-AI companies that rely on massive, heterogeneous training sets.

Midjourney has already pushed back in earlier cases, arguing that training on large collections of images and producing novel outputs is transformative and protected by fair use — the same line other AI companies have taken. The company has characterized its service more like a search engine or a creative assistant, with users ultimately responsible for how they prompt the model. That defense, however, has yet to be fully tested in court, and the recent flurry of complaints means judges will soon confront thorny questions about copying, transformation and the limits of copyright in the age of generative systems.

Why studios say this matters

From the studios’ perspective, the risk isn’t only lost licensing fees. WBD and others warn that high-fidelity, easily downloadable copies of iconic characters can dilute their brands, undercut merchandising and licensing deals, and confuse consumers about authenticity — an especially acute worry for companies that make billions from character franchises across movies, TV, toys and theme parks. The complaint frames the issue as both economic and cultural: these fictional characters are the backbone of studio businesses, and studios argue the law should protect the market they’ve built around them.

What Midjourney and the industry say (so far)

Midjourney’s public posture has been defiant in prior suits: the company has stressed that generative systems remix and transform, that many images used for training are publicly available, and that it enforces terms of service intended to curb blatant piracy. The company and some AI advocates argue that strict liability for outputs could chill innovation and make it practically impossible to develop creative AI tools without licensing every image on the internet — a logistical and financial nightmare. Those arguments will now be weighed against the studios’ claim that Midjourney’s outputs are direct, unauthorized derivatives of specific copyrighted works.

Industry groups and trade bodies will likely weigh in. The Motion Picture Association and other defenders of the creative industries have already signaled concern about unchecked copying by AI, arguing that consumer confusion and economic harm to rights holders justify legal remedies. Meanwhile, AI research and civil-liberties groups warn against broad rulings that could hobble new tools or entrench incumbents’ control over culture. Expect amici briefs and high-profile commentary as the case proceeds.

What to watch next

Legally, the case raises a few immediate battlegrounds:

• Training data: Did Midjourney ingest unauthorized copies of copyrighted works when it trained its model? The answer may hinge on discovery and whether the company retained logs or training corpora that can be tied to specific images.
• Output similarity: How close do the generated images have to be to the originals to count as infringement? The studios are using side-by-side comparisons; Midjourney will argue about transformation and novelty.
• User responsibility vs. platform liability: If users prompt the model to generate copyrighted characters, is the platform liable or the user? Courts will look to precedent around online platforms, but AI introduces new wrinkles.

Practically, the case could result in anything from a narrow injunction and damages to broader rules about how commercial AI models are trained and policed. If Warner Bros. Discovery wins meaningful relief, other companies may follow, pushing AI labs toward licensing deals or stricter content-filtering systems. If Midjourney and similar firms prevail, the decision could embolden more permissive AI development and make licensing the exception rather than the rule.

For artists, illustrators and IP holders, the suits are a signal: courts may provide new tools to protect character art and derivative markets — but outcomes are uncertain and will take time. For casual users and hobbyists, the immediate change may be that certain prompts become blocked or that platforms add clearer warnings about copyrighted content. For businesses building on generative AI, the message is simple and urgent: legal clarity is coming, and it could reshape product design, content moderation and licensing strategies across the industry.

Warner Bros. Discovery’s lawsuit against Midjourney crystallizes the collision course between century-old copyright law and cutting-edge generative technology. It’s another heavyweight entry in a litigation wave that will force courts to reckon with whether — and how — copyrighted culture can be used to train and fuel systems that reproduce it. Expect a long, technical, and publicly watched fight: the outcome won’t just affect Midjourney and Warner; it could set precedents for how the internet itself is remixed for the next generation of AI.