Google says users can explore a generated environment and change a running scene with written instructions: altering the weather, moving the camera or adding a character. The model also generates ambient sound, music and dialogue.
Google plans subsequent access through Project Genie for AI Ultra subscribers and selected Gemini API and AI Studio partners. It has announced no general release timetable or Genie 4-specific price. The proposed uses, including interactive creation and training robots or agents, remain applications put forward by the laboratory.