A few months ago, OpenAI made a big splash with Sora. This new AI tool can create detailed and lifelike videos with just a few text prompts. While that tool is still in the private testing stage, Google decided to go ahead and announce its own AI-based video generator.
During the 2024 Google I/O developers conference, the company announced Veo, from its Deepmind AI division. In a blog post today, Google stated:
Veo generates high-quality 1080p resolution videos in a wide range of cinematic and visual styles that can go beyond a minute. With an advanced understanding of natural language and visual semantics, it generates video that closely represents a user’s creative vision — accurately capturing a prompt’s tone and rendering details in longer prompts.
Google added that the model behind Veo can understand specific film and cinematic terms when a user types them in, such as “timelapse” or “aerial shots of a landscape”.
Google DeepMind posted a (slightly over) one-minute video made with Veo on its YouTube account today. It shows footage of a neon-lighted city from above before it cuts into a car running through that city, The video ends with the car and the city switching over to a more photo-realistic clip. The video also shows the text prompts that were used to make the example Veo clip.
In another video, Google shows how actor, writer and director Donald Glover is collaborating with Google DeepMind to use Veo as the basis for an upcoming short film.
Google plans to make Veo"s feature available to a select group of users who sign up to access its VideoFX tool. Interested people can sign up for the waitlist now.
At the moment there"s no work on when Veo will be made generally available. However, Google did add it is working on adding Veo features to creators of YouTube Shorts videos.