Google DeepMind تعلن Genie 4: عوالم تفاعلية تغيّرها بالكلام، ولها صوت وحوارGoogle DeepMind announces Genie 4: interactive worlds you change with words, with sound and dialogue

Google DeepMind has announced Genie 4, a "world model" that generates interactive environments you move through in real time and can change with words. You type something like change the weather, move the camera or add a new person, and the scene changes in front of you. You get around with standard keyboard controls.

What is new is that it is DeepMind's first real-time world model to generate audio natively: wind in the trees, waves on the beach, music and even dialogue between characters that adapts to what you do. DeepMind says the biggest leap is how it simulates people, with characters that move naturally and react to their surroundings. The model builds on the Gemini family, including Gemini 3.8 Flash, which watches the scene and breaks goals into steps.

Trusted testers in creative industries, journalism, education, gaming and engineering have tried it. Filmmakers can run several takes of a scene in real time, and game developers can turn an idea into an explorable prototype in seconds. All video and audio it generates carries SynthID, a hidden watermark that helps identify AI-generated content.

But it is still an experimental research preview for trusted testers only, coming later to Google AI Ultra subscribers in Project Genie and early access partners in the Gemini API and Google AI Studio. DeepMind admits scene details can drift in longer sessions, and that rendering text, audio accuracy and complex physical interactions are still being researched.