Google Introduces Genie 4 AI to Build Interactive Virtual Worlds from Text Prompts
· GlobeWire Daily Editorial Desk · Source: www.indiatoday.in

Image: India Today · Photo source · Resized and letterboxed to 16:9; no AI generation
Google DeepMind has introduced Genie 4, an update designed to generate interactive 3D virtual worlds from text prompts. Released as a research preview on Tuesday, the model allows users to explore and alter photorealistic spaces in real time without stopping the simulation.
Key Features and Capabilities
- Real-Time Edits: Users can type changes mid-simulation, such as altering camera angles, adding rain, or introducing new characters.
- Live Audio: The AI generates synchronized background noise, music, and reactive digital character speech.
- Extended Runtimes: Simulations can run for up to two minutes, doubling the duration of the previous Genie 3 model.
- Gemini Integration: Google's Gemini AI acts as a director to plan logic and steps, while Genie handles visual simulation.
Practical Applications and Testing
Google is testing the technology for robotics, allowing companies to simulate tasks safely before real-world deployment. Google representatives suggest potential use cases in gaming, interactive learning, filmmaking for rapid storyboarding, and creative asset development alongside tools like Veo and Gemini Omni.
Current Limitations and Availability
Developers noted that long-running simulations can experience "drift," where objects lose detail or disappear. Audio synchronization and rendering readable text inside the worlds remain under development.
Genie 4 is currently available to a small group of trusted testers. Google plans to roll it out in the coming weeks to Google AI Ultra subscribers through Project Genie at $100 per month, as well as to early access partners via the Gemini API and Google AI Studio.
AI-assisted reporting; source claims are attributed and have not been independently verified.



