Google DeepMind Releases Genie 4 World Model With Real-Time Editing and Audio
aiinsider ·
Google DeepMind has released Genie 4, its latest world model, in a research preview phase. This new system generates highly photorealistic and interactive simulated environments that allow users to modify scenes in real-time using text prompts. Users can alter weather conditions, adjust camera angles, or introduce characters and objects while the model simultaneously produces matching audio, including music and dialogue. The simulation maintains scene consistency for up to two minutes, which is double the duration of its predecessor, Genie 3. Gemini acts as an integrated director to plan steps and verify frame quality, enhancing the coherence of the output. In robotics experiments, this integration allowed for real-time simulation of task breakdowns, offering a potential testing ground for robot deployment before physical implementation. Although early testers find value in film prototyping, engineers note that challenges such as object drift and audio accuracy remain areas for improvement.
aiinsider