Disney Unveils AI-Powered Star Wars Anzellan Animatronics With AWS
Walt Disney Imagineering and Amazon Web Services built three AI-driven Star Wars characters that react to guests in real time.
Walt Disney Imagineering, the company’s theme park design and research arm, has unveiled three AI-powered Star Wars Audio-Animatronics characters built with Amazon Web Services (AWS), the cloud computing unit of Amazon. The figures, called Anzellans, respond to guests and to each other in real time, according to coverage published Wednesday by Laughing Place and Wilson’s Media. The Hollywood Reporter and CNET also published reports on the project, according to the headlines that accompanied them.
What happened
The three characters are named Vilo Vele, Kado Bre’a and Jibu Bre’a. They belong to the same species as Babu Frik, the tiny droid technician from the film “Star Wars: The Rise of Skywalker.” An earlier version appeared at the D23 fan event in August, Laughing Place reported.
The characters sit together in a crate, can move and look around, and interact with one another without a puppeteer. They have onboard cameras, microphones and sensors, and they speak a mix of English and the fictional Anzellan language, Laughing Place reported.
Kyle Laughlin, senior vice president of Walt Disney Imagineering R&D and Technology and Engineering, called them “our most advanced Audio-Animatronics in the world,” according to Wilson’s Media.
How the technology works
Wilson’s Media reported that the figures share a robotic character platform with Olaf, the free-roaming snowman robot, and the BDX droids that have appeared at Disney parks. The AI layer comes from AWS and uses Amazon Bedrock, a service that gives developers access to multiple AI models, including vision, language and world models. Because Bedrock is not tied to one model, Disney can swap models to improve speed, cost or performance without rebuilding the experience, the outlet reported. AWS also said it used Kiro, its AI coding agent, to turn Imagineering’s creative briefs into working software.
Disney says people, not AI, control the content. According to Wilson’s Media, Imagineers wrote every line of dialogue, which is prerecorded, and designed each character’s personality and movements. The AI chooses which responses and animations fit a given moment, for example reacting to a guest’s birthday button.
Samira Panah Bakhtiar, general manager for Telco, Media, Entertainment, Games and Sports at AWS, said in comments carried by Laughing Place: “What makes this innovation truly special is that it exemplifies physical AI.”
On privacy, Wilson’s Media reported that data gathered during an interaction is used only for that experience. It is not stored in the cloud for Disney to reuse or used to train AI models.
Where you can see them
The Anzellans will appear at AWS re:Invent, Amazon’s annual cloud conference, in Las Vegas. Both outlets said they will be at the Customer Showcase at Caesars Forum from Nov. 30 to Dec. 4. Some headlines describe the debut as coming “later in November,” but the conference dates reported by the outlets begin on Nov. 30.
What is not yet known
Disney has not announced when or where the characters will reach a theme park. Laughlin said guests would likely meet them indoors first, according to Wilson’s Media, and he declined to give specifics. He described the characters as “an experiential merchandise platform,” suggesting interactions could happen in shops rather than at traditional meet-and-greets.
Laughlin also said the effort is “not about short-term playtests,” and that he hopes to see “an incredible pace of scale” over the next 10 years, per Wilson’s Media. He said any such experience could eventually feature characters from any Disney franchise. Neither outlet reported pricing, a rollout schedule or the number of figures planned for parks.


