Google Project Genie
Google Project Genie offers a rare glimpse at where interactive AI worlds are headed, it shows how world models could reshape games, simulation, and storytelling. Built on Google DeepMind’s Genie 3 world model, Project Genie lets users create, explore, and remix interactive environments generated in real time from simple text prompts or images.
Instead of loading a fixed 3D scene, Project Genie generates the world as you move through it. Walk forward and the path ahead forms on the fly. Fly, drive, or switch perspectives, and the environment responds instantly. That sense of motion driven creation is what makes this prototype stand out, even in its rough early state.
World sketching is the starting point. Users describe a setting, character, and point of view, then fine tune a preview image before stepping inside. From there, exploration feels playful and experimental, with worlds that lean more artistic than realistic. Remixing existing worlds and downloading short videos adds another creative layer. Watch the impressive video below.
Instead of loading a fixed 3D scene, Project Genie generates the world as you move through it. Walk forward and the path ahead forms on the fly. Fly, drive, or switch perspectives, and the environment responds instantly. That sense of motion driven creation is what makes this prototype stand out, even in its rough early state.
World sketching is the starting point. Users describe a setting, character, and point of view, then fine tune a preview image before stepping inside. From there, exploration feels playful and experimental, with worlds that lean more artistic than realistic. Remixing existing worlds and downloading short videos adds another creative layer. Watch the impressive video below.
