Google just launched Flow, an AI filmmaking tool powered by its advanced Veo 3 video model, Imagen image generator, and Gemini AI, aiming to let creators generate cinematic video complete with native audio directly from text prompts. The company pitches Flow as an intuitive way for storytellers and advertisers alike to explore ideas without bounds.
Lights, camera, algorithm: At its heart, Flow is designed as a creative playground where Google says Veo 3 excels in prompt adherence and realistic outputs, while Gemini handles conversational prompting. Users can import assets or use Imagen to generate “ingredients,” theoretically ensuring consistency across scenes and even allowing native audio like environmental sounds and dialogue to be spun directly by the AI.
Ticket to create: Flow offers features like “Camera Controls” for shot mastery, “Scenebuilder” for extending clips, and “Asset Management.” Inspiration comes via “Flow TV,” a showcase of AI-generated content. Access began May 20th for Google AI Pro and Ultra subscribers in the U.S., with Ultra offering higher limits and Veo 3’s full audio capabilities.
Beta blues and deepfakes: Early hands-on exploration, notably by The Verge, has revealed Veo 3’s sophisticated and sometimes unsettling power, with instances of AI-invented dialogue and convincingly fabricated scenes. While some guardrails exist, visualizing plausible fictions is strikingly easy. Testers also found that extending clips with Scenebuilder currently means losing Veo 3’s native audio, as the process seemingly defaults to the older Veo 2 model.
Future of film or fancy filters: Google is betting big on AI transforming video creation, empowering a “new wave of creators.” Yet, early limitations and the ease of generating convincing fabrications highlight a learning curve for both the tech and its users.
