Google has taken a significant step in generative artificial intelligence with the launch of Project Genie, an experimental system that can create interactive 3D virtual worlds from simple text and image prompts — a development that could eventually reshape how digital environments are built.
The tool, now available in early access to US-based subscribers of Google’s premium AI Ultra plan, moves generative AI beyond static outputs such as text, images and video, into real-time, explorable environments. Developed by Google DeepMind, Project Genie is powered by Genie 3, a so-called “world model” capable of dynamically simulating environments as users navigate them.
In a statement, Google said trusted testers had already used the underlying model to generate a wide range of immersive worlds and experiences, prompting the company to expand access through a dedicated interactive prototype focused on world creation.
Unlike established game engines such as Unity or Unreal Engine — which rely on manually built assets, complex scripting, physics tuning and lengthy iteration cycles — Project Genie generates landscapes, structures and scenery on the fly. A single prompt can produce a navigable environment, collapsing what traditionally takes teams of developers weeks or months into seconds.
The web-based prototype allows users to sketch worlds through descriptions, explore them by walking or flying, and remix existing environments. Users can also export short video clips of their sessions. Current experiences run at around 720p resolution and 24 frames per second, with individual interactions limited to roughly one minute.
Google, however, has been careful to stress that Project Genie remains a research prototype. The company acknowledged limitations, including imperfect realism, occasional inconsistencies with prompts, basic physics, reduced character control and latency issues. World generation is also capped at 60 seconds per session.
Access to the tool is currently restricted to Google AI Ultra subscribers, a plan priced at $250 per month, which includes higher AI usage limits, 30 terabytes of cloud storage and access to advanced coding tools.
While far from production-ready, Project Genie offers a glimpse into a future where creating interactive digital worlds may no longer require deep technical expertise. For developers, designers and the gaming industry at large, Google’s experiment signals a potential shift — from painstaking manual world-building to AI-generated environments shaped by imagination and prompts alone.
