Google DeepMind introduced Genie 4 on Tuesday. This AI tool makes worlds on a computer, and people can move through them. Google says these worlds look more real than before, including the people and the way things move. The tool also makes music, background sounds and speech.
Users can type instructions to change a world while they explore it. They can change the weather, move the camera or add people and objects. Another Google AI tool, Gemini 3.8 Flash, helps plan what happens next and checks the pictures. Genie 4 can keep a scene the same for about two minutes, up from one minute with Genie 3.
Google says the tool could help people plan films, try game ideas and create places for robots to practise tasks. But objects can move to the wrong place or act in strange ways during longer visits. The sounds are not always right, and words inside these worlds can be hard to read.
Only a small group of trusted testers can use Genie 4 now. Google plans to offer it to Google AI Ultra customers through Project Genie in the coming weeks. Some early partners will also get access through other Google tools.
