Google, in collaboration with Donald Glover, has showcased an AI video generator called Veo AI to compete with OpenAI’s Sora. Although there’s no set launch date or rollout plan yet, the demo indicates that It can create high-quality, realistic videos similar to Sora. Check this also OpenAI Sora : Text-to-Video Revolution (2024 Update)
What’s great about Veo AI? “You can make mistakes faster” Glover said in a video during Google’s I/O 2024 livestream. “In art, that’s what you want – to quickly learn from your mistakes.”
Speaking at Google I/O in Hawaii, Google DeepMind CEO Demis Hassabis said, “This AI creates high-quality 1080p videos from text, image, and video prompts.”
Veo AI
Veo AI and Sora both offer the same resolution and video length, around one minute. Since neither product is released yet, it’s hard to compare them in detail. However, Hassabis mentioned that It will let users “further edit your videos using additional prompts,” a feature Sora currently lacks according to early users. Read this also for How to Spot AI Generated Images and Videos 2024?
It’s unclear what this AI was trained on. About a month ago, YouTube CEO Neal Mohan told Bloomberg that using YouTube videos to train Sora would violate YouTube’s terms of service. However, Alphabet, which owns YouTube, also owns Google, the creator of Veo AI.
Mohan suggested in that interview that YouTube content is used to train Google’s AI models, but only with user consent.
We do know that Veo AI is built on Google and Deepmind’s past projects, such as the Generative Query Network (GQN) from 2018, last year’s VideoPoet, Google’s early video generator Phenaki, and Lumiere, which was demoed earlier this year.
Donald Glover’s specific AI filmmaking project hasn’t been revealed yet. In the video shown at Google I/O, Glover mentioned he’s been interested in AI for a few years and reached out to Google, not the other way around. He stated, “We got in contact with some of the people at Google and they had been working on something of their own, so we’re all meeting.”
How does it work?
Google Veo AI creates videos from text prompts by understanding the described scene. It identifies visual elements like objects and actions to form a mental image.
Veo AI then generates a video with the right camera angles, lighting, and colors. Users can customize their videos and keep frames consistent. Veo also follows responsible practices to prevent misuse and ensure privacy and fairness.
To use Google Veo AI, you need to join a waitlist. Here’s how:
- Visit the AI Test Kitchen and find the Veo waitlist.
- Fill out the form and submit it.
- Wait for your invitation.
Once you have access to Google Veo AI and VideoFX:
- Describe the scene you want to create.
- Customize if needed.
- Let Veo AI generate the video.
- Review and refine the video.
- Download or share your video.
Explore additional features for more options.
Currently, the public can’t try Veo , but there is a waitlist signup page available.
What a fantastic blog layout! For what duration have you been blogging? You make it look so easy. Both the general design and substance of your website are excellent.