Google's New AI Lets You Build Playable Worlds in Seconds

Imagine describing a magical forest or snapping a photo of your backyard, then watching it transform into an interactive world you can explore with a keyboard. That's exactly what Google's Project Genie does, and it just became available to more people.

The technology, called a world model, creates video environments that respond to your movements in real time. You can walk around using keyboard controls, and the AI generates what comes next as you explore. It feels like playing a video game, except the entire world is being invented on the spot.

Google first showed off this technology last year as Genie 3, but kept it limited to a small group of testers. Now they've polished it up and connected it to their latest AI models, making the experience smoother and more reliable.

Getting started is surprisingly simple. You can either upload a photo for reference or just describe what you want in words. The AI first creates a still image based on your input, which you can tweak if needed. Once you're happy with it, Genie transforms that image into a 60-second explorable world running at 24 frames per second.

The system even lets you remix Google's pre-made worlds with different characters and visual styles. Every time you run the same prompt, you get slightly different results since the AI generates everything fresh. You can download videos of your explorations to keep or share.

The Ripple Effect

While Project Genie is still a research prototype with some limitations like input lag and the 60-second time limit, it represents something bigger than just a cool tech demo. Game designers and educators are already imagining uses for rapidly prototyping interactive experiences. Teachers could create custom learning environments. Storytellers could sketch out scenes before committing to full production.

The technology isn't perfect yet. Sometimes the physics don't behave quite right, and there will be occasional glitches in how worlds look or respond. Google is also still figuring out content policies after initially allowing recreations of copyrighted games, then blocking some of those prompts.

Right now, Project Genie requires Google's AI Ultra subscription at $250 per month, which makes sense given how much computing power it takes to generate interactive video in near real time. But Google has confirmed plans to expand access over time as the technology becomes more efficient.

The company stresses this is still experimental, but that's part of what makes it exciting. We're watching the early days of technology that could eventually let anyone sketch interactive ideas as easily as drawing on paper.