Gemini Generates Videos with Your Digital Clone—A Journalist Finds It Eerily Unsettling

A Wired journalist created a video where his digital clone tells a story with the facial expressions and gestures of the original. The video was made using a new feature in the Google Gemini app. The result turned out so realistic that the author felt somewhat unsettled.

How the Digital Clone Works

Google has integrated into its AI assistant Gemini a tool that allows generating videos with a digital avatar of the user. The process is simple: you upload your photo or video sample, and the system creates a virtual clone ready to tell any stories. The voice is synthesized based on a speech sample, and facial expressions and gestures automatically adjust to the script text.

The feature is positioned as a tool for content creators, marketers, and businesses. Instead of shooting video yourself, you can upload a script, and the avatar will turn it into a finished video. There's no need for a studio, lighting, or camera operator. The process takes minutes, and the quality is comparable to professional video.

Eerily Accurate Results

When the Wired journalist first saw his digital clone, the effect was unexpectedly strong. The avatar copies not just physical appearance, but microexpressions: head tilts, fleeting facial expressions, speech intonations, even habitual gestures. All of this is captured and reproduced with such accuracy that it creates an uncanny valley effect—when the copy is so similar to the original that it seems wrong.

Seeing yourself, but not quite yourself—it's strange and slightly unsettling.

At first glance, the video looks like an ordinary, live video of a real person. It's hard to immediately notice the synthesis. This raises a whole range of moral questions: who owns this avatar, can it be stolen or used for malicious purposes, how accurately can the digital clone impersonate the original online.

Google's Vision

Google views digital avatars as one of the major trends of the coming decade. The company sees them as the future of content creation: everyone will be able to delegate video creation to their digital clone. This saves time, money, and resources.

For businesses, the advantages are clear:

• Fast creation of video content without complex production
• Ability to reshoot video in seconds by changing only the script
• Scaling: one person can appear in hundreds of videos
• Integration into apps and platforms

For now, Google is being cautious. The feature requires explicit user consent, and the company doesn't allow creating avatars of famous people without permission. But the technology is developing, and regulation is lagging behind.

What This Means

Digital clones are moving from labs into mainstream products. This opens enormous opportunities for productivity and creativity, but simultaneously creates new risks: deepfake videos, identity theft, mass misinformation—all these threats become more acute as the technology improves.