The legal risks surrounding generative AI have intensified as Disney and Universal filed a joint copyright lawsuit against image-generation platform Midjourney. The complaint, submitted this week, alleges that the company’s technology enables users to generate unauthorized reproductions of well-known characters from the studios’ extensive catalogues—including properties owned by Marvel, Lucasfilm, DreamWorks, and more.
According to the 110-page filing, Midjourney acts as “a virtual vending machine” that produces close approximations of protected characters, such as Darth Vader, Shrek, Iron Man, and Bart Simpson. The studios claim this capability stems from Midjourney’s unauthorized scraping of copyrighted materials during the training of its algorithms. The suit argues that these outputs—generated even from vague prompts—demonstrate clear evidence of intellectual property infringement.
What makes this particular legal action stand out is the visual documentation included in the complaint: side-by-side comparisons of Midjourney’s outputs with original stills from film and television productions. These illustrations aim to present the case not as a complex debate about machine learning or abstract fair use principles, but as a straightforward instance of plagiarism in the digital age.
The studios also criticize Midjourney’s selective content moderation policies, which ban certain political or explicit content while allowing the generation of images based on characters still protected under copyright law. The lawsuit notes that despite receiving a cease-and-desist from Disney legal counsel, Midjourney allegedly continued operations without any public course correction or negotiations.
The broader legal context is rapidly evolving. Courts have begun issuing early rulings that challenge the fair use defenses often cited by AI companies. Recent decisions involving companies like Thomson Reuters and Meta suggest that courts are increasingly skeptical of models trained on protected data without licenses or consent. These rulings, while not binding precedent in this specific case, add weight to Disney and Universal’s claims and may shape how future litigation unfolds.
Midjourney, which launched in 2022 and initially operated through Discord before transitioning to a standalone platform, has gained popularity for its subscription-based generative art services. Prices start at $10 per month, attracting artists and hobbyists interested in AI-driven creativity. However, the legal backlash now calls into question whether its outputs cross the line from inspiration into replication.
At the heart of the lawsuit is a familiar debate: whether generative AI tools are simply remixing publicly available culture or profiting from the unlicensed reproduction of original work. Under current copyright law, fictional characters are entitled to protection beyond their source material—a protection that Disney and Universal argue is being actively disregarded by Midjourney’s model.
The studios are seeking financial damages, restitution of profits, and a permanent injunction to halt any further use or generation of their intellectual properties. Midjourney has not issued a public response as of this writing.
The outcome of this case could significantly influence the boundaries of generative AI, particularly when it comes to how platforms are trained and the limits of permissible output. As the entertainment industry tightens its stance on unauthorized use of IP in AI systems, companies in the AI space may face increasing pressure to revisit how their tools are built—and what legal frameworks they’ll need to operate within going forward.
