Disney and Adobe have signed a multiyear agreement to develop private generative artificial intelligence models trained on Disney’s intellectual property, extending their work across theme parks, film development and ESPN. The partnership will use Adobe’s Firefly Foundry platform to build tools tailored to Disney’s creative assets and production needs.
The arrangement centers on company-specific models rather than general-purpose AI services. The models will be private and owned exclusively by Disney, giving the entertainment company greater control over systems designed to generate material using its characters, designs and visual styles. The financial value of the agreement has not been disclosed.
Initial work will support Disney Experiences, the business encompassing its theme parks, resorts and cruise operations, alongside concept development at Walt Disney Studios’ StudioLAB innovation hub. Adobe will also work with ESPN on tools intended to improve engagement with sports fans and make its content archive more accessible.
Walt Disney Imagineering, which designs and develops attractions and immersive experiences, has already introduced a customized image-generation model and a tool that transforms hand-drawn sketches into more developed concept art. Those applications give designers a way to examine different visual approaches before moving further into a project.
Kyle Laughlin, Disney’s senior vice president of technology, engineering and research and development, said the tools have shortened some creative processes from months to hours or days. He said faster iteration can help teams identify potential changes earlier, rather than making expensive revisions during construction or implementation. That assessment concerns creative development work, not the full timetable for building an attraction.
Laughlin also said the tools are intended to support Imagineering employees rather than replace them. The stated purpose is to accelerate experimentation and reduce the time spent developing successive versions of an idea, while leaving creative specialists involved in the process.
For a business built around valuable entertainment properties, the use of proprietary training material is a central feature of the agreement. Copyright disputes surrounding generative AI often concern whether protected works can be used to train models without permission. A model developed for an intellectual property owner using its own assets presents a different authorization structure from a public service drawing on material supplied by unrelated rights holders.
Private access and ownership of a model do not, by themselves, settle every legal question surrounding its outputs. Copyright protection for AI-assisted work can depend on the extent of human authorship, while rights in underlying material remain relevant to how generated content is used. The reported terms of the Disney-Adobe agreement do not detail provisions governing those issues, liability or the handling of particular categories of creative assets.
The planned applications also differ by division. Imagineering is already using image tools for design exploration, while StudioLAB will examine their role in developing visual concepts. At ESPN, the collaboration will focus on audience engagement and access to archived content. No detailed rollout schedule for those broader applications has been disclosed.
Sources: Business Upturn, Українські Національні Новини (УНН), Finimize