We expected Google to go hard on AI this year at Google I/O, and the tech giant hasn’t disappointed at all. One of the biggest new developments, at least for filmmakers, is Flow, a new AI video tool that blends together several different models from Google DeepMind into one package.
Google calls Flow an “evolution of VideoFX,” an experiment the company launched in Google Labs last year. It brings everything from Google’s top advanced models, such as Imagen, Veo, and Gemini, into one place.
Google says the combination of these three models in one tool allows for unprecedented levels of prompt adherence. The inclusion of Imagen here allows you to bring your assets directly into Flow, so you can create characters using text-to-image generation.
One of the biggest advancements that Flow brings to the table is scene consistency. Despite AI video tools getting better and better, consistency is still a tough thing to sort out sometimes. Google claims that Flow will make the process easier.
Flow will also give users full camera control, so they can change the motion, angles, and even perspectives of the shot. The Scenebuilder functionality will let you seamlessly edit and extend existing shots.
Over the past several months, we’ve seen some very promising AI video tools, including Runway Gen-4, which creates some pretty stellar shots. Whether Flow and the AI models that power it will carry the same weight remains to be seen.
So far, though, some of the demo videos that Google has released have looked very promising.
On top of launching Flow, Google also revealed info about new updates to its most advanced models with the arrival of Veo 3 and Imagen 4. Google claims these are the strongest models it has released for generative media.
Veo 3 is the first time that Google’s AI video model has allowed you to create audio for your clips, too, so that should be a big improvement for filmmakers working with it. The company is also expanding access for Lyria 2, its music-focused generative AI tool.
Flow, Veo 3, and Imagen 4 are all available today. Flow will be available for Google AI Pro and Google AI Ultra plans in the US, with plans for more country support soon. US-based Ultra subscribers will have access to Veo 3 starting today. Vertex AI enterprise subscribers can also access it today.
Imagen 4 is the most widely available of these models, and it will be available today for the Gemini app, Whisk, Vertex AI, and across Slides, Vids, Docs, and other features in Google Workspace.
