- Google AI image-making tools are now blocking many Disney character prompts
- The shift follows a cease-and-desist letter from Disney
- Users who previously used Gemini and Nano Banana to generate images of characters like Yoda and Elsa are being stonewalled
Disney has pushed back against AI making unauthorized images of its characters, at least using Google's tools. After a December cease-and-desist letter from Disney accusing Google of allowing its models to behave like an unlicensed “virtual vending machine” for intellectual property, the spigot of Gemini and Nano Banana-produced images of Disney characters has cut off entirely.
Prompts that easily produced high-quality images of familiar characters are now greeted with denial messages explaining that the request cannot be fulfilled due to "concerns from third-party content providers." Google's tightening of its filters in response to Disney is another sign that the Wild West era of AI images may be coming to a close.
Of course, few AI prompt blocks are truly airtight. Users have shared some limited loopholes already, such as uploading a photo of Buzz Lightyear and combining it with the figurine prompt to make a virtual toy of the character. But enforcing clean copyright boundaries across AI systems that combine text and images will likely never be perfect.
Still, the 32-page legal warning to Google from Disney has done its work. You won't be making Iron Man, Darth Vader, or Elsa from an eight-word text prompt anymore. Disney argued that this amounted to unauthorized reproduction of copyrighted works at scale and that Google's models had been trained on copyrighted Disney material without permission.
The company demanded that Google immediately halt the production of unlicensed character imagery, stop training the models on its intellectual property, and put in place systems that prevent future unauthorized use. It also stated that Disney had been raising concerns for months without meaningful progress.
AI copyrighting
Google publicly assured Disney and other IP holders that it was working on better copyright controls, like Content ID, which give rights holders some ability to manage how their content is used. The company also pushed back on the training accusation, firmly asserting that its models are trained on public web data. Of course, any random Disney characters and details may be scooped up in that net. Nonetheless, the legal pressure seems to have prompted policy changes.
Disney didn't mention it in its letter, but the fact that the entertainment giant has its own AI plans almost certainly influenced the legal effort. Disney recently signed a $1 billion licensing deal with OpenAI, authorizing the use of Disney characters within OpenAI's video generator Sora.
Disney clearly sees economic value in generative AI, but wants to make sure it has its hands on the wheel. Users may eventually find themselves navigating a fragmented landscape, where each AI platform has different rules about what can be generated depending on which rights holders are in active negotiations.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!
And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.
