Google has introduced Flow, its latest AI-generated software that can create movies from user prompts and inputs. This eliminates the need for expensive real actors, film sets, and locations while speeding up the process of creating entertaining productions.
Flow leverages several of Google's advanced AI models - Veo for video generation, Imagen for image generation, and Gemini for tasks and prompts. The three models work together to come up with realistic images and scenes based on user prompts.
The filmmaker can start by typing prompts to create images that best represent each scene. When they are happy with the look, users can add additional prompts detailing the actor movements to generate realistic action shots. Multiple shots can be combined with transitions for a professional look.
The system maintains object consistency throughout different shots, so an actor dressed in a certain way in one shot will look the same in another. Camera controls allow virtual filmmakers to position the camera and plan camera moves using common cinematography terms, such as pan, tilt, and dolly. All of the prompts and scenes can be organized for easy re-use.
Creatives can come up with new ideas using Flow TV, a catalog of visual examples generated with Veo that details the exact prompts and methods used. The ability to iterate through many ideas allows filmmakers to rapidly turn their story concepts into cinematic entertainment.
Use of Google Flow requires a subscription of $19.99 per month to Google AI Pro or $249.99 per three months to Google AI Ultra.