Disney Star Wars animatronics use AWS AI to talk with guests
Walt Disney Imagineering R&D has shown three Anzellan prototypes that read a room with hidden cameras and microphones, then respond through AI models on Amazon's cloud. Every line they speak is still written by Disney Imagineers.

Key takeaways
- Walt Disney Imagineering R&D has revealed three Audio-Animatronic prototypes named Vilo Vele, Kado Bre'a and Jibu Bre'a.
- Hidden cameras and microphones feed several AI models running on Amazon's AWS cloud servers.
- A fast-decision model covers quick reactions such as eye contact, while larger models plan longer interactions.
- All dialogue and movements are prewritten by Disney Imagineers, so the characters never stray from the script.
In this article
Three Anzellans head to Las Vegas
Disney Star Wars animatronics that react to visitors in real time have been revealed by Walt Disney Imagineering R&D. The three prototypes use Amazon's AWS cloud and AI models to respond to what guests say and do, according to The Verge. They are due to make their first public appearance in Las Vegas later this month, at the AWS re:Invent conference.
The machines are named Vilo Vele, Kado Bre'a and Jibu Bre'a. Each one is based on the Anzellans, the small alien characters seen in The Mandalorian and Grogu, a film released earlier this year. The Verge reports that the three units are built with distinct personalities, so no two behave in quite the same way.
How the robots sense a guest

Most animatronics in Disney's parks either repeat a fixed set of moves, or depend on a puppeteer working out of sight to make the character seem spontaneous. The Anzellans take a different route. Each unit carries hidden cameras and microphones, which capture a visitor's movements and speech.
That material is passed to several AI models running on AWS servers rather than inside the figure itself. One is a fast-decision model that produces immediate reactions, such as turning quickly to make eye contact so the character appears to be paying attention. Larger models handle the slower work of shaping a longer exchange with a guest.
The script still belongs to Imagineers
The AI does not invent what the characters say or do. According to The Verge, every movement and every line of dialogue that the models select and trigger is written in advance by Disney Imagineers. The aim is to keep the animation faithful to each character, including the individual quirks that give the three Anzellans their own identities.
In practice, that means the models choose the timing and the response, but the Anzellans never improvise beyond the material prepared for them. The Verge draws a contrast with Disney's Olaf robot, which was shown recently and relies on an operator using a Steam Deck gaming handheld to bring the character to life. No such hidden operator is involved with the Anzellans.
What this could mean for the parks
Disney has not confirmed that the Anzellans will become a permanent attraction. The Verge reports that the characters, and the technology behind them, could eventually turn up in Star Wars: Galaxy's Edge and at other Disney parks. No date or location for a wider rollout has been given.
For readers who follow theme park technology, the interesting part is the division of labour. Cloud-based AI handles perception and timing, while human writers keep control of character and dialogue. The only confirmed public outing so far is the Las Vegas conference, where AWS attendees will be able to interact with the three prototypes.
Sources
Spotted a mistake? See our corrections policy. Read how News by OnDevice works in our editorial policy.



