Google Reveals New AI Model for Generating Interactive Virtual Worlds

In depth

Google DeepMind has unveiled Genie 4, a model that generates interactive virtual environments in real time. According to Bizportal, the model generates environments with background sounds, music and dialogue, and relies on Gemini. According to Ynet, it is capable of creating characters, sounds and physics rules. According to Channel 14, the environments include realistic physics and built-in sound, and can be controlled with a keyboard.

According to Bizportal, the model serves as a training tool for robots and agents, and is used for game development. According to Ynet, the technology may change the way robots are trained and games are developed, but it is limited to two minutes of simulation. In the first stage the model is open to selected testers; according to Bizportal, it will later reach Google AI Ultra subscribers, and according to Channel 14, access will open exclusively to subscribers of the elite package.