Google says its AI-powered video tool Veo can capture tone, simulate real physics, and understand filmmaking terms like "timelapse."
Google is squaring off against OpenAI on a new front: AI-powered video creation.
Google's new text-to-video tool called Veo can generate 1080p videos based on text, image, and video prompts, the company announced on Tuesday from the Google I/O stage.
Google showed off some impressive clips created by Veo in the announcement.
In one, a crocheted elephant ran through the savannah; another showed an aerial shot of a lighthouse on a cliff with waves crashing beneath; in a third clip, a person's face is covered in smoke and then disappears.
Veo can capture tone and simulate real physics, Google said, and also understands filmmaking terms like "timelapse."
"We're exploring features like storyboarding and generating longer scenes," DeepMind CEO Demis Hassabis said onstage Tuesday. "Veo gives you unprecedented creative control."
Google announced it has invited artists to experiment with Veo.
"Over the past year, we've made incredible progress in enhancing the quality of our generative media technologies," said Google vice president of product management Eli Collins and senior research director Doug Eck in a blog post.
"We've been working closely with the creative community to explore how generative AI can best support the creative process, and to make sure our AI tools are as useful as possible at each stage," the pair wrote.
Veo has invited a group of creators to the product — and actor and director Donald Glover and his creative studio, Gilga, experimented with the tech.
"Everybody's going to become a director and everyone should be a director," Glover said in a video shown at Google I/O. "Because at the heart of all this is just storytelling. The closer we are to being able to tell each other our stories, the more we'll understand each other."
There's waitlist for Veo open now, Google said.
The emphasis on working with creatives comes amid criticism from entertainment industry vets, who previously raged against AI tools as a danger to their professions and art after OpenAI revealed its video generator Sora.
Filmmaker Tyler Perry saw the technology as a threat while director Tim Burton previously said AI-created imitations of his work lacked "humanity" and a "soul."
Still, some early testers thought the tool could boost their creative process and noted that Sora still required human oversight.
Veo will live within VideoFX, Google's recorder app, and some features will arrive in YouTube Shorts and other Google products in the future, Google said.
To distinguish that they are AI-generated, all videos created in Veo will tout a new SynthID — an imperceptible digital watermark developed by Google.
Google said it's been working on generative video for years, and Veo builds on previous ventures like Lumiere, VideoPoet, WALT, and Phenaki.
Now it's revealed Veo — just months after OpenAI shocked with its own AI-powered video creation tool.
OpenAI revealed Sora in February with eerily accurate scenes of people walking through cities and mammoths trundling through the snow.
Google seemed to take a shot at OpenAI's tool in its announcement on Tuesday, noting that Veo can create clips over a minute long; Sora's videos are limited to under a minute.
