I Cloned Myself With Gemini’s AI Avatar: The Result Was Unsettlingly Accurate

It’s a warm, sun-drenched afternoon at Dolores Park in San Francisco, and I’m belting out a birthday tune for a cartoon dinosaur. As I finish the song, a cupcake topped with a pink candle materializes in my empty hand. When I blow out the flame, the CGI creature beams with a serene, satisfied expression.

The man in this AI-generated video looks and sounds exactly like me, but it’s not real. It was created using one of the latest features in Google’s Gemini app: avatars. These digital doppelgängers function much like the core features of OpenAI’s now-shuttered Sora app. Essentially, it’s a digital clone of yourself that can be inserted into AI-generated videos. Powered by Google’s new Omni video model, this feature is exclusive to paying subscribers.

I’m on Google’s AI Pro plan, which costs $20 a month, and I quickly hit the usage cap (it resets every five hours). After asking just a few questions and generating two 10-second clips featuring my avatar, I was told to come back later.

Video: Reece Rogers

My first two glimpses of Omni’s capabilities with my likeness were me singing to a dinosaur in San Francisco and surfing under the Golden Gate Bridge. I felt a mix of awe and unease. The content was undeniably cheesy, with awkward moments and nonsensical outfits, but that was undeniably me. I zoomed in on its face with my fingers, studying the mouth movement. The teeth were slightly off, but otherwise, it was Reece , right down to the chin fat.

Unlike OpenAI, which once let users decide whether others could generate AI videos using their likeness, Google restricts this feature so that only adult users can create videos with their own avatar.

Setting up my avatar through the Gemini app took about five minutes. I sat in a well-lit room, pointed my phone’s camera at my face, and read a string of two-digit numbers. Then I slowly turned my head to the right and then to the left. That was it. Reece 2.0 was born and ready to star in my deepfake adventures. (Be mindful of what you’re wearing during this process, since your outfit will likely appear in the AI generations , more on that later.)

Let’s break down the birthday clip frame by frame to fully unpack my reaction. The full prompt was: Generate a video of me singing the happy birthday song to an aging dinosaur at the top of the hill at Dolores Park.

AI-generated clip by Reece Rogers

The first second features a millennial pause, because even AI Reece has some ingrained habits. What’s most striking initially is the photorealistic setting. Rather than placing my avatar on some generic hill in a random park, the background of Google’s AI video is remarkably similar to the actual location. From the palm tree-lined sidewalks to the looming Salesforce Tower in the distance, it’s immediately clear which park this is, even though the output isn’t flawless. It makes sense that a company known for mapping the planet could pull this off.

As AI me started to sing , with a less pitchy baritone than I can actually manage , the first few bars felt natural. I bounced my hands up and down on the beat, like a mini conductor. Then I stutter on the word “to,” and Gemini cuts to a wider-angle shot as the real chaos begins. A vanilla cupcake appears out of nowhere, and I exhale a cloud of smoke to blow out the celebration candle. (Honestly, how rude of AI Reece. It’s not your special day.)