Google DeepMind’s Genie 4 lets users steer simulated scenes in real time with words—and generates native audio as the world changes.

📰 Announced October 27, 2026, Genie 4 can change camera movement, weather and characters while generating ambient sounds, music and interactive dialogue. Rollout has started with limited trusted testers. Google AI Ultra access in Project Genie and early partner access through the Gemini application programming interface (API) and AI Studio are coming soon.

🫵 For a filmmaker with trusted-tester access, the proposed workflow is to try a different camera move or weather condition inside an ongoing simulation, with sound generated alongside it—exploring scene alternatives before a shoot.

📈 Google DeepMind’s interactive world model could give Google AI Ultra subscribers another reason to keep paying if scene steering and dynamically generated audio become useful tools for recurring creative work. The promised subscription access is still forthcoming, however, and the limited tester rollout does not show whether the simulations are reliable enough for production workflows. Robot and agent training remains a potential application rather than demonstrated cost savings.