Google Tests Gemini Avatars With 3D Head Scans

Evidence is mounting that a 3D “Avatar” feature is nearing release inside Gemini, giving you a fast way to scan your face and drop a realistic version of yourself into AI‑generated images and XR experiences. Fresh references found in a recent Google app build point to phone‑based head scanning, a web creation flow for desktops, and a simple prompt tag that inserts your digital self on demand.

What the New Clues Show About Gemini Avatar Features

Text strings discovered in version 17.11.54.sa.arm64 of the Google app repeatedly mention Gemini “Avatars.” Launching the flow reportedly triggers an intro screen that asks you to capture short video with your phone’s camera to scan your head. One prominent instruction reads, “Add yourself to any frame you create by typing @me or @%s into your prompt,” which strongly suggests one‑tap self‑insertion into generative scenes. Notably, Google support language still leans on the older “likeness” phrasing, hinting at a transition in branding rather than a brand‑new concept.

From Likeness to Avatar: How Google Is Rebranding the Tech

“Likeness” debuted on Android XR to solve a practical problem: video calls while wearing a headset. Instead of appearing in a call with a visor strapped to your face, a photoreal proxy represents you. Now, the same idea appears to be broadening into Gemini’s creative tools under the more approachable “Avatar” name. Earlier internal builds even referenced “Characters,” underscoring that Google has been refining both the tech and the branding as it converges XR presence with generative media.

How Scanning Likely Works Across Phones and the Web

The capture flow resembles consumer photogrammetry: a short, multi‑angle video produces a textured 3D mesh of your head that Gemini can render into new scenes. Modern pipelines often blend classic structure‑from‑motion with neural rendering techniques inspired by work like NeRFs, enabling robust results from a standard RGB camera. That mirrors how other platforms approach identity in immersive spaces—Apple’s Vision Pro uses Personas for video calls, while Meta has explored high‑fidelity “Codec Avatars” in research. The discovery that creation can occur in a browser signals desktop support using a webcam or uploaded footage, not just phones.

What You Could Do With It in Creative and XR Workflows

For creators, the ability to type “@me” turns time‑consuming compositing into a near‑instant step. Imagine generating a product mockup, marketing scene, or social thumbnail and placing your true‑to‑life face and head pose right into the frame—no green screen, no manual masking. In XR, the same avatar can represent you during calls, collaborative whiteboarding, or virtual events, aiming for continuity across chat, canvas, and headset. The result is a single identity that travels with you through generative content and immersive apps.

Privacy and Safety Questions Around Scanning Your Face

Face scans are sensitive. The critical questions are where the data is processed, how long it is stored, and what controls you have. Best practice would allow on‑device creation where feasible, clear deletion options, and explicit off‑switches for third‑party access. Watermarking will also matter if avatars are embedded in synthetic content; Google has promoted SynthID for identifying AI‑generated media, and similar safeguards could help flag avatar‑assisted outputs. Guidance from bodies like NIST underscores the importance of consent, revocation, and protection against spoofing when biometrics are involved.

How Close Is Launch for Gemini Avatars and Head Scans

References inside shipping app builds usually indicate that user testing is underway and server‑side flags could enable trials without a full update. Coupled with a web‑based workflow and existing XR “Likeness” foundations, the practical groundwork appears largely in place. Expect a staged rollout—likely starting with select phones and regions, then expanding to desktops and XR devices—as creation quality, latency, and safety reviews are tuned.

Why It Matters for XR and AI in Generative Identity

High‑fidelity avatars are the connective tissue between today’s generative tools and tomorrow’s spatial computing. They make collaboration more human, speed up branded and creator content, and help immersive platforms feel less like demo tech and more like daily software. As major players converge on standards like OpenXR and invest in volumetric presence—from enterprise training to telehealth—the ability to reliably scan, personalize, and reuse your identity across surfaces could be the feature that nudges XR from novelty to habit.

The takeaway: the branding may still be settling, but the signals are clear. Gemini Avatars with quick head scanning and simple prompt insertion are close enough that you should start thinking about how you’ll use them—and how you’ll protect them.