Technology Popular

Gemini will build an avatar from your face and your voice, and let other people use it

You record a face and a voice once, then generate images and video without filming again. Permission can be granted for a set period and revoked, but content already made survives.

Gemini will build an avatar from your face and your voice, and let other people use it

Google has added personal avatars to Gemini. The feature builds a digital double of a user from their own face and voice, and that double can then appear in images and video without the person ever standing in front of a camera again.

The setup is deliberately short. You photograph your face and record your voice. If you begin on a computer, Gemini displays a QR code; you scan it with a phone or tablet and record the voice there, because a handset microphone is a more predictable thing than whatever is attached to a desktop.

What you can do with it

Once the avatar exists, it can be used to generate a range of pictures and video. The point of the feature is that nothing needs reshooting: the likeness is stored, so each new piece of content is a generation rather than a recording session.

Making video with a personal avatar requires a Google AI plan. The announcement did not state the price, and it did not say which countries have the feature, how long a generated video may run, or how many a plan allows.

Revoke, but not erase
Gemini lets you withdraw permission for your avatar at any time. Anything already made with it stays made. Graphic: Tech BD.

Lending your face to someone else

The part that deserves attention is sharing. An avatar can be handed to friends, family or colleagues by sending an invitation to their Gmail address. Once they accept, they can use your likeness in content they generate themselves.

The owner keeps a measure of control. Gemini’s settings let you decide who may use the avatar and for how long, granting permission for a fixed period, and you can cancel at any moment. After cancellation that person can no longer produce new images or video with your face.

The sentence to read twice

Content already made does not disappear. Revoking permission stops future generation; it does nothing to anything generated while the permission was live.

That is a reasonable engineering decision — Google cannot reach into other people’s downloads and files — but it changes what “revoke” means in practice. Lending your likeness is not like lending a key you can take back. It is closer to lending a photocopier: you can stop them making more copies, but every copy made up to that point is theirs.

For a colleague putting your face into a team presentation, that distinction is harmless. For a relationship that ends badly, or a colleague who leaves on bad terms, it is the whole question.

Why this is arriving now

Avatar generation has existed in various products for a while, usually as a novelty. What Google has added is the social layer: a permissions system for faces, with invitations, expiry dates and a revoke button. That is an admission that the hard problem was never making a convincing double — it was deciding who is allowed to use one.

It also sets a precedent. If lending your likeness becomes an ordinary thing people do at work, the norm will be established by whatever default these settings ship with, long before anyone writes a rule about it.

Before you record

Two things are worth settling first. Decide whether you want the avatar shareable at all, because the safest configuration is one you never lend. And if you do lend it, set an expiry rather than leaving it open-ended — a permission with an end date expires on its own, while an open one survives until somebody remembers to end it.

Source: Prothom Alo

Written by

Tech BD

Tech BD is an independent technology publication based in Dhaka, published in English and Bangla. Articles appear under the masthead rather than under individual bylines; corrections and source questions reach the editor at info@tech-bd.com.