Google Vids now lets you star in your own AI videos
TechCrunch Sarah Perez ● Covered by 2 sources
Google just gave Vids a glow-up: you can now clone yourself into AI videos with a selfie and a voice clip. While OpenAI's Sora went dark, Google's betting big on the same idea, just bolted onto Workspace.
Based on reporting by TechCrunch, Sarah Perez — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Google announced Thursday that its Vids tool is getting a substantial upgrade, and it's a telling one. Users will soon be able to build a digital avatar of themselves using nothing more than a selfie and a short voice recording. That avatar can then star in videos generated through prompts, turning what used to be a fairly buttoned-up presentation tool into something closer to a personal AI film studio.
The bigger change under the hood is the arrival of Gemini Omni, Google's multi-modal model, inside Vids. Omni takes a written prompt plus whatever reference images you feed it and blends them into a finished video. It's not just for generating clips from scratch, either. Omni can clean up lighting on footage shot on a phone, swap out a distracting background, or layer in effects. And Google added step-by-step editing, so a change to your video no longer means starting the whole thing over.
What's notable is the timing. This lands right as OpenAI's Sora, the tool that made AI avatar videos a mainstream curiosity, has shut down. Google seems to be reading that as an opening rather than a warning sign. By folding these features into Google Workspace, the company is clearly aiming Vids at business use first, things like internal updates or training videos, the kind of content that rarely goes viral but gets made constantly inside companies.
Still, personalized avatars and conversational editing push Vids into territory currently occupied by dedicated AI video startups like HeyGen, Synthesia, Captions, and D-ID. That's a crowded field, and Google's advantage is less about flashy tech and more about distribution: Vids already sits inside Workspace, where plenty of people already live for work.
Google is also drawing some careful lines around the avatar feature. Each avatar is tied to the account holder's own likeness and Google account, and every video gets an invisible SynthID watermark. Access is restricted to users 18 and older in select regions for now. It's a pointed contrast to Sora's early days, when Sam Altman reportedly let users generate strange videos of him using his own likeness. Don't expect anyone making mischief with Sundar Pichai's face through this tool anytime soon.
My take — AI-written commentary, not fact-checked reporting
Tying avatars strictly to your own verified likeness and account is the boring, sensible choice, and boring is exactly what this category needs after Sora's free-for-all phase. Google clearly isn't chasing viral chaos here, it's chasing the Workspace paycheck, betting that companies want tidy training videos more than they want weird internet novelty. That's a much less exciting pitch than Sora ever made, but it's probably the one that survives.
Read more about this at: TechCrunch