Google's generative AI filmmaking program Flow has reached a milestone. The tech giant confirmed exclusively to CNET that Flow creators have produced over 100 million AI videos in the program. Thanks in part to its advanced AI video model, Veo 3, Flow allows users to generate video clips and edit them together to create scenes.
It's been 90 days since Google surprised us with Flow at its annual I/O developers conference. According to Elias Roman, senior director of product management for Flow in Google Labs, much of the time since has been spent "hustling just to keep up with the demand."
Flow is a departure from Google's previous generative AI work. For years, the company's AI efforts have been focused on Gemini, its all-in-one chatbot. It has flooded its products with AI, like with Search's AI overviews and Gmail's AI-generated summaries. Its research assistant tool, NotebookLM, with its AI audio generator that can transform documents into personal podcasts, continually rolls out new features.
The industry leader has spent billions of dollars trying to win the race to develop the most advanced AI for average Google searchers, developers and, yes, even artists and creators. 100 million AI videos is a significant milestone for the company, and it helps show us what the future of AI-enabled creation might look like.
Getting in the AI Flow
To compete with Midjourney and Stable Diffusion, Google created a crop of AI image models, originally named ImageFX and now known as Imagen (pronounced "imagine"). Its previous generative media models were better suited for amateur or enthusiast creators, not professionals, and they didn't dominate the AI creative space. That all changed with Veo 3.
Google dropped Veo 3, its newest AI video model, at May's I/O conference. Veo 3 leapfrogged the competition with a somewhat obvious but first-in-industry advancement: AI videos with synchronized, AI-generated audio. The model garnered a ton of attention online, and Google reported over 40 million AI videos just seven weeks later.
"What Veo 3 allowed was a much wider set of people to create very compelling videos, engaging all the senses out of the box. You didn't have to stitch together a toolkit," said Roman. "To be able to do the Foley [ambient sounds], the sound effects, the soundtrack, the dialog, all of that, and not make the user think about each of those modalities in a specific way, I think, is a big unlock, too."
Veo 3 is one of several AI models you can use in the filmmaker tool. Flow was built for professional creators and filmmakers, a step beyond simple image and video generation available with Gemini. Google intentionally moved away from its original ImageFX nomenclature and built off its interface, Roman said, and wanted Flow to combine the most advanced Imagen and Veo models with Gemini, which was used in the training of Veo and "basically speaks native Veo."
Flow is one way to combine all those AI models and pieces, uniting Google's different generative AI models for seamless video creation and editing.
... continue reading