Google’s newest XR story is not a headset launch, a benchmark or a promise that a chatbot can write a film. At Venice Immersive, its 100 ZEROS team backed three works that put the technology in different supporting roles: a character you can speak with, a natural-history guide you can question, and a conventional teaser given spatial depth.
That distinction is the point. The best idea in this small slate is not that AI appears on screen. It is that interaction has to earn its place in the story.
The three projects premiered during the 83rd Venice International Film Festival, which runs September 2–12. Google says all are intended for Android XR; the festival’s own listing confirms that NEVATARS is a 31-minute mixed-reality work in its official Venice Immersive selection. Independent coverage from ANSA and Screen Daily also places both NEVATARS and Galápagos: The Last Eden in the program.
Three experiments, three jobs for the technology
| Project | What Google says it adds | What is confirmed now |
|---|---|---|
| NEVATARS | Gemini and Vertex AI-powered conversational moments, alongside real-time avatar work | A 31-minute mixed-reality installation by Andy Serkis and Josh Nelson Youssef, in Venice Immersive’s official selection |
| Galápagos: The Last Eden | A Gemini-powered digital Charles Darwin within an immersive wildlife journey | Narrated by Margot Robbie and shown at Venice Immersive; The Immersive Wire describes it as an interactive Android XR documentary |
| Sedona | Google’s 2D-to-3D XR auto-spatialization for a psychological-thriller teaser | Google says the teaser is available on YouTube; it is not presented as a consumer XR release |
The difference between those rows matters. NEVATARS makes conversation part of the drama: its official festival synopsis follows an artist facing the discarded characters in a forgotten computer folder. A responsive character can serve that premise because the work is already about what happens when created things acquire a voice.
Galápagos takes the more familiar documentary route. A digital Darwin offers a plausible interface for curiosity—ask a question, follow an animal, learn why an ecosystem works. That can be more engaging than turning a narrator into an all-purpose assistant, but it also puts a high burden on the experience: a wrong, dull or overly scripted answer breaks the spell immediately.
Then there is Sedona, the most restrained use of the technology. Google describes it as a teaser whose presentation gains depth from 2D-to-3D auto-spatialization, rather than an invitation to converse with a character. It is a useful reminder that spatial computing does not always need a new interaction model. Sometimes it needs a better sense of scale and framing.
Venice is proof of intent, not a distribution plan
This is where the announcement needs a little cold water. Festival installations are ideal places to test novel cameras, displays, interaction patterns and audience patience. They are not proof that those ideas have reached a repeatable consumer format.
Google Italia’s more detailed briefing says NEVATARS was designed as a 31-minute smart-glasses installation for XREAL Aura glasses, which it describes as not yet on sale. Google has not announced public release dates, prices or availability for NEVATARS or Galápagos. The responsible reading is therefore simple: they are Venice experiences, not products you can buy or download today.
That does not diminish the creative work. It clarifies the test. Spatial storytelling will not become ordinary because a large platform can fund prestigious projects. It becomes ordinary when someone can discover a work, use hardware they already own, experience it without venue staff, and recommend it to a friend without a long explanation.
Google has been building the platform side of that equation, while Android’s easier password-manager transfers show the less glamorous work required to make a device ecosystem livable. The same lesson applies here: a great XR installation is only the beginning. Discovery, compatibility, comfort, access and a clear reason to return decide whether it can travel.
Why this is a better AI use case than a talking prop
The term “AI film” has become a shortcut for a confusing number of things: generated visuals, production tools, synthetic performers, personalization and interactive dialogue. Venice gives Google a chance to separate them.
The strongest of the three propositions is not “talk to a famous historical figure.” It is using a responsive system where an audience question changes the texture of a scene without claiming the system is the author. Google says its 100 ZEROS initiative works with filmmakers to explore emerging technology while keeping human creativity central. That is a corporate ambition, not an independently verified outcome—but the projects’ distinct formats at least show an attempt to make the tool subordinate to the work.
There is a caveat. The reporting and program listings establish the projects, their creators and their festival context; they do not independently test the quality, accuracy or guardrails of the Gemini-powered interactions. Until public access exists, claims about how natural these conversations feel should stay claims.
Bottom line
Google’s Venice slate is a promising signal because it does not pretend every XR story needs the same AI trick. NEVATARS uses interaction to make creativity personal; Galápagos uses it to make exploration active; Sedona suggests depth can matter even when nobody talks back.
But viewers should resist mistaking an elegant festival premiere for a finished market. The real win will come when these experiences have a route out of Venice—and a reason to exist after the headset comes off.