Code for Dignity Privacy
← All essays

Consent and deepfakes

Who owns your face in the age of AI?

What if the most intimate thing about you is no longer your heartbeat, your memory, or your name but your face?

By Lianna Adams ·

In the age of AI, a face can be copied without touch, sold without permission, and revived long after the person is gone. We are entering a world where identity is no longer only lived; it is also extracted, simulated, and circulated.

Who owns your face in the age of AI?

In older worlds, a face was a sign of presence. It told you someone was there: alive, singular, accountable. Now a face can become a file, a dataset, a prompt, a synthetic performance. The same image that once marked the boundary of a self can now be turned into a product, a deepfake, a memorial, or a weapon. That is the quiet rupture AI has introduced: it does not merely imitate us; it makes imitation feel normal.

The legal world is starting to notice. Watermarking, provenance, and disclosure rules are gaining attention because synthetic content is getting harder to distinguish from human-made content. But the deeper problem is not technical detection. It is moral permission. A system can recognise a face and still fail to recognise a person.

Consent is not enough if it can be forgotten

We are told consent solves everything. But consent in the digital world is often a checkbox, a buried setting, or a one-time yes that gets stretched into endless future uses. A photo shared casually can become training data. A voice note can become a clone. A selfie can become a likeness asset that travels farther than the person who took it.

That is why the phrase “my face” matters so much. It sounds simple, but it contains a philosophy. My face is not just an image of me. It is one of the ways I appear to the world as myself. When AI copies it, the issue is not only ownership. It is whether the copy still carries the dignity of the original.

Likeness as property, likeness as self

One response is to treat likeness like property. That approach has appeal because it gives people a right to license, restrict, and revoke uses of their image and voice. It also fits a world where faces can generate value and synthetic media can circulate like currency.

But property is not the whole story. A chair can be owned without moral strain. A face is different. It is not merely something a person has; it is part of how a person is known. That is why “owning” a face is both useful and inadequate. Ownership suggests control. Personhood suggests something deeper: that the face belongs to the self in a way that should not be reduced to market logic.

The better word may be stewardship. We should think less about possessing a face and more about protecting a boundary. A boundary says: before you copy me, you must ask. Before you reuse me, you must justify. Before you turn me into a model, you must honour the fact that I am not raw material.

The dead do not disappear

AI also changes the meaning of death. In the past, death ended the person’s ability to speak for themselves. But now a face can keep working after death, impersonating the dead with unnerving realism. That raises a hard question: if a person no longer lives, can their likeness still be violated?

It should be able to be violated, because the harm is real. Families can be wounded by posthumous deepfakes. Public memory can be rewritten by synthetic endorsements. The dead may not feel harm in a biological sense, but the living do. And the dead left values, wishes, and relationships behind them. A civilisation serious about dignity should not treat death as a free license for reuse.

This is where posthumous consent becomes essential. If someone says in life, “Do not recreate me,” that request should matter. Not because the dead need entertainment rights, but because consent should not become meaningless the moment a body stops breathing.

A child’s face should not begin as a default asset

The deepest injustice may be generational. Children are growing up in a world where every stage of development can be photographed, shared, stored, and used to train systems they never chose. Before they have a voice, they already have a digital trace. Before they can consent, they can be copied.

That is why your “digital birthright” idea is so powerful. It imagines that a child enters the world with built-in likeness protection, not as a luxury, but as a basic right. A future worth wanting would not make people fight to reclaim themselves after being harvested. It would begin by assuming their face is theirs, and only theirs, unless they decide otherwise later.

The practical version of this is not mystical. It is policy, design, and etiquette. Opt-in by default. Clear revocation. Portable consent. Stronger protections for minors. And social norms that treat asking as the beginning of respect, not the inconvenience before it.

What AI really threatens

The true danger is not that AI can copy us. It is that it can make the copy feel more convenient than the original. Once that happens, the human becomes optional in the imagination of the system. A face becomes a resource. A voice becomes a template. A person becomes a profile with permissions attached.

That is why this question matters so much now: who owns your face in the age of AI? The answer should not be “the platform,” and it should not be “the model.” It should be you, but not in the shallow sense of consumer control. You should own your face because your face is one of the places where your humanity becomes visible.

In the end, the argument is not only legal. It is philosophical. A society reveals what it believes by what it allows machines to do with the human form. If we allow faces to become endlessly extractable, we will not only lose privacy. We will train ourselves to see persons as content.

And once that happens, the theft is no longer just of images. It is of the idea that a human being is more than the sum of their copies.

So the real question is not whether AI can reproduce a face. It is whether our culture will still recognise that a face is a human claim before it is a commercial one. If we fail there, we do not just lose control of images; we weaken the meaning of consent itself.

Originally published in The Human Strata. Republished with permission from Lianna Adams.