Disney has sent Google a cease-and-desist letter ordering the company to stop using its AI systems to create images and videos featuring Disney-owned characters, a letter obtained by Variety states. The movie studio and TV network group accuses Google’s Gemini of playing a part in piracy that is “on a massive scale,” and is demanding technical, as well as policy, changes overnight.
The dispute underscores a key battle line in the generative AI age: who gets to control and profit from how famous people are represented when machines can create them at will, and under what licensing terms.
What Disney wants from Google regarding AI character use
Disney’s letter demands that Google block Gemini from creating content featuring characters across Disney’s portfolio, which includes Pixar, Marvel, and Star Wars. Reporters and users have demonstrated that Gemini can display pictures of Mickey Mouse, Black Panther, or Darth Vader upon request, so long as prompts do not otherwise violate platform rules.
Cease-and-desist letters are not enforceable as orders of the court. They are a formal notice, explaining what an alleged offender is accused of doing wrong and putting the recipient on record. Should Google fail to act, Disney may even threaten litigation based on an unauthorized derivative work and trademark claims involving the commercial exploitation of its own IP.
A licensing deal with OpenAI alters the context here
The move comes as part of a broader licensing deal between Disney and OpenAI that will allow ChatGPT Images and Sora, in accordance with company statements and industry reporting, to create officially licensed imagery featuring Disney characters. Also of note in the reporting: Disney is acquiring an equity stake in OpenAI, which is worth $1 billion or higher, meaning it sees strategic value in betting on AI as “the future,” one that falls within a corporate strategy to keep its IP locked down in how it appears across different platforms.
The juxtaposition is striking: a licensed pipeline into another leading AI service, on one hand, and a push to lock down similar output elsewhere. For Disney, consistency matters. Without a tight leash, a flood of off-brand or misleading content can dilute character value and confuse consumer trust.
The risk for AI outputs in the court of law
There are several avenues available to rights holders under copyright and trademark law. Characters are protected works; illegal copying and look-alikes may constitute derivative works. By the same token, trademarks layer in a further category — one where use that is not properly authorized and connotes sponsorship or confuses consumers can create liability even apart from classic copyright-type assertions.
The U.S. Copyright Office has made it clear that works of authorship produced entirely by AI cannot be afforded copyright protection, but that doesn’t authorize the repurposing of others’ IP. The agency’s concurrent investigation into AI and copyright, meanwhile, has elicited vast amounts of public and industry comments, underscoring unanswered questions about training data, output controls, and fair use limits.
For platforms, the problem isn’t just user behavior but whether product design foreseeably facilitates infringement at scale. That’s why many AI companies rely on fine-grained safety filters, style-block lists, and even watermarking systems. Google, for example, has shown off SynthID watermarking for AI images, yet watermarks don’t cover whether an image’s portrayal should even have been created in the first place.
How Google could respond to Disney’s cease-and-desist
In terms of implementation, Google could base its blocks on characters, require proof of license for particular prompts, or by default provide generic look-and-feel outputs (thereby circumventing protected elements). Other competitors have layered “copyright shield” policies and strong prompt classifiers to keep brand and character misuse at lower levels, with specifics of these systems differing.
Another likely move would be more explicit user messaging when prompts call for trademarked or copyrighted characters. Open refusals, accompanied by safer alternatives like creating new characters, can help to grease the wheels while still respecting rights. Disney’s letter received no immediate comment from Google.
Why it matters for creators and users as AI scales
AI art fans often use favorite characters without permission for noncommercial fun, but platforms operate at mass scale, and results can easily stray into commercial or reputational damage. Brands like Disney, meanwhile, see character integrity as a key asset; managing their presentation and context is one way they safeguard billion-dollar franchises and long-term licensing revenue.
The immediate question is whether Google will modify Gemini’s outputs to meet Disney’s wishes. The bigger takeaway is clear: as AI image and video tools grow up, licenses — not permissive defaults — are the way to carve out high-profile IP. Rights holders are running far ahead in defining the rules, and the platforms that can move quickly will determine how iconic figures exist in the synthetic media age.
