Google I/O 2024’s keynote session was a 112-minute-long affair the place the corporate made a number of main bulletins centered on artificial intelligence (AI). The bulletins ranged from new AI fashions to integration of AI into Google merchandise, however maybe one of the crucial fascinating introductions was Veo, an AI-powered video era mannequin, that may generate 1080p decision movies. The tech big mentioned that the AI instrument can generate movies that transcend the one-minute mark. Notably, OpenAI additionally unveiled its video AI mannequin dubbed Sora in February.

In the course of the occasion, Demis Hassabis, co-founder and CEO of Google DeepMind, unveiled Veo. Saying the AI mannequin, he mentioned, “At this time, I am excited to announce our latest and most succesful generative video mannequin referred to as Veo. Veo creates high-quality 1080p movies from textual content, picture and video prompts. It could actually seize the small print of your directions in several visible and cinematic types.”

The tech big claims that Veo can carefully comply with prompts to grasp the nuance and tone of a phrase after which generate a video to resemble it. The AI mannequin can generate movies in several types like timelapse, close-ups, fast-tracking pictures, aerial pictures, and varied lighting and depth of area pictures. Aside from video era, the AI mannequin can even edit movies when the person offers it with an preliminary video and a immediate so as to add or take away one thing. Additional, it could additionally generate movies past the one-minute mark both via a single immediate or by way of a number of sequential prompts.

To unravel the issue of consistency in video era fashions, Veo makes use of latent diffusion transformers. This helps in decreasing the cases of characters, objects, or the whole scene flickering, leaping, or morphing unexpectedly between frames. Google highlighted that movies created by Veo will probably be watermarked utilizing SynthID, the corporate’s in-house instrument for watermarking and figuring out AI-generated content material. The mannequin will quickly be obtainable for choose creators by way of the VideoFX instrument at Google Labs.

Veo’s similarities with OpenAI’s Sora

Whereas neither of the AI fashions is offered to the general public but, each share a number of similarities. Veo can generate 1080p movies for a period that may surpass one minute, whereas OpenAI’s Sora can generate movies of as much as 60 seconds. Each fashions can generate movies from textual content prompts, photos, and movies. Based mostly on diffusion fashions, each are able to producing movies from a number of pictures, types, and cinematography tehcniques. Each Sora and Veo additionally include AI-generated content material labels. Sora makes use of the Coalition for Content material Provenance and Authenticity (C2PA) normal whereas Veo makes use of its native SynthID.