Google has unveiled a new tool aimed at simplifying the creation of AI-generated videos, named Flow. This announcement comes alongside the introduction of the Veo 3 video generation model, enhancements for the Veo 2 model, and the launch of the Imagen 4 image generation model.
Flow allows users to create eight-second AI-generated video clips using text-to-video prompts or by providing images that help guide the model in understanding user preferences. Additionally, it features scenebuilder tools that facilitate the combination of multiple clips into a cohesive video.
While Flow functions similarly to a film editing application tailored for AI video production, it serves as a quick tool for filmmakers to bring their creative visions to life. A demonstration by Thomas Iljic, a product manager at Google Labs, showcased several examples of Flow’s capabilities.
In one instance, we observed a video where the camera zoomed out to reveal it was playing on a TV, followed by a wider shot of the room before the camera seamlessly transitioned outside to view a passing truck.
The demonstration appeared to flow seamlessly, although my perspective was limited to a small window on Google Meet, leaving some uncertainty regarding any potential AI anomalies that might not be immediately noticeable. Flow is designed to help filmmakers efficiently generate their concepts rather than focus on longer video productions, according to Iljic.
Regarding the newly announced models, Veo 3 promises enhanced video quality, simplified prompting capabilities, and the ability to generate both video and audio, including dialogue. Matthieu Lorrain, the creative lead at Google DeepMind, explained that Veo 3 has improved sensitivity to longer prompts and can effectively handle sequences of events within those prompts.
The Veo 2 model will provide features such as camera controls and object removal. Meanwhile, Imagen 4’s advancements include improved quality, broader format options for exports, and a notable increase in its ability to produce coherent text instead of nonsensical AI-generated phrases.
Flow is now available for users in the United States who subscribe to Google’s new Google AI Pro and Google AI Ultra plans. The Google AI Pro plan offers essential Flow features with a limit of 100 generations per month, while the Google AI Ultra plan provides maximum usage limits and early access to Veo 3, inclusive of native audio generation, according to an official blog post.
