If generative AI is the new printing press, your face and voice just became the most valuable “content” on earth.

Over the last two years, you’ve probably seen deepfake clips of politicians saying things they never said, creators discovering AI clones of their voices on sketchy apps, or influencers finding their faces pasted onto explicit content. Until recently, the legal answer to “Can they do that?” was often a fuzzy “it depends.”

Denmark is trying to turn that “it depends” into a very clear “only if you say so.”

The Danish government has proposed something radical in its simplicity: treat your personal characteristics, like your appearance and voice, as if they were a kind of copyrightable work. If the proposal becomes law, Danish citizens would effectively gain copyright-style control over their digital likeness, including how AI can imitate and distribute it online.AP News That idea is already being talked about as a “Denmark model” for the rest of the world.

What Is the “Denmark Model” in Plain Language?

The shorthand “Denmark model” describes a legal shift with three big ingredients:

  • Deepfakes of real people are explicitly targeted.
  • A person’s likeness (face, voice, other recognizable traits) is treated as a protected asset.
  • You get a clear right to say who can imitate and share your likeness – and to demand takedowns when they do it without consent.

In 2025, Denmark announced plans to change its copyright law so that sharing AI-generated deepfakes that imitate someone’s appearance or voice without consent could be banned. If passed, this would give people a copyright-like right over their likeness, not just over traditional creative works like photos or songs.AP News

This goes further than just saying “deepfake porn is illegal” or “harassment is bad.” It reframes the whole thing as: your likeness is a work, and you are the rightsholder. If someone uses AI to generate or distribute “you” without your authorization, they are infringing.

That sounds a lot like what celebrities and athletes already do with endorsement deals. Denmark’s twist is aiming to generalize that logic to everyone, not just public figures.

How Denmark Got Here: From Privacy to Personality Rights

This didn’t come out of nowhere. Denmark already has:

  • Strong privacy and data rules under GDPR, enforced by its Data Protection Agency (Datatilsynet). Photos of identifiable people are treated as personal data, which generally means you need a legal basis – like consent or legitimate interest – to publish them online.Datatilsynet FAQ
  • Personality rights under its Penal Code that protect you from certain unauthorized uses of your image and name, especially in commercial contexts.Overview of personality rights

But generative AI changes the game. It’s no longer just about a photographer uploading a picture of you; it’s about anyone, anywhere, spinning up thousands of convincing fake videos of “you” with a single selfie or short voice clip.

That is why Denmark’s 2025 initiative zoomed in on deepfakes and “personal characteristics” – appearance, voice, and other traits – as things that should be explicitly protected through copyright-like rules, not only through privacy or harassment law.European Parliament briefing on the Danish approach

You can think of it as stacking protections:

  • Privacy law: “You can’t just process and publish my personal data any way you like.”
  • Personality rights: “You can’t exploit my image or name to sell stuff without permission.”
  • Denmark model: “You also can’t use AI to fabricate me and spread that content as if it’s real, because my likeness itself is protected.”

Denmark isn’t doing this in a vacuum. It sits inside the bigger machinery of the EU AI Act, the world’s first comprehensive AI law, which entered into force in 2024.EU AI Act overview

For your digital likeness, three parts of the AI Act matter most:

  1. Deepfake labelling (Article 50)
    The AI Act requires that AI-generated or AI-manipulated content that “appreciably resembles” real people or events and could falsely appear to be authentic – in other words, deepfakes – must be clearly disclosed as such. That includes synthetic images, audio, and video.AI Act text

  2. Biometric and emotion recognition transparency
    If you are being exposed to biometric categorization (e.g., systems trying to infer your age, gender, or mood from your face), you must be informed.AI Act transparency FAQs

  3. Watermarking and machine-readable signals
    Providers of generative AI systems must ensure their outputs are detectable as AI-generated, for example through watermarking or other machine-readable markers, so platforms and regulators can distinguish synthetic content at scale.AI content watermarking overview

The EU Act is, in essence, about transparency and risk management. It says: label the deepfake, make it detectable, and inform people when AI is analyzing or imitating them.

The Denmark model adds a sharper edge: it is not just that deepfakes must be labelled – many of them may simply be illegal to share at all if they imitate a real person’s likeness without permission.

Why This Matters to You (Even if You Are Not Famous)

You might be thinking: “I’m not a politician, actor, or influencer. Why should I care?”

Two reasons:

  1. Generative AI doesn’t care who you are.
    Tools like ChatGPT, Claude, and Gemini are mostly text-based today, but multimodal models (text + image + audio + video) are rapidly mainstream. Consumer apps that can clone a voice from a 10-second voicemail or build a face-swapped video from a single selfie already exist. The barrier is not fame; it’s data.

  2. Your risk is reputational and relational, not just commercial.
    The most damaging deepfakes so far have often been non-consensual sexual images, harassment campaigns, or scams targeting ordinary people, not just celebrities. The Denmark model treats your appearance and voice as something only you can license, which makes it easier to demand takedowns and, potentially, compensation.

In practical terms, if something like the Denmark model were adopted where you live, it could mean:

  • Platforms having clearer obligations to remove AI-generated content that uses your likeness without consent.
  • AI tools needing proof of consent (or a licence) to legally train on, or generate, your face and voice.
  • A legal basis for you to negotiate actual licensing deals if a company wants to use a digital double of you in an ad, a game, or a virtual experience.

How This Interacts With AI Tools You Actually Use

You interact with generative AI every time you:

  • Use ChatGPT, Claude, or Gemini to draft content.
  • Upload a selfie to an AI portrait generator.
  • Try a voice-cloning app that reads your newsletter in “your” voice.
  • Use a video tool that animates a still image of you into a talking avatar.

Under EU rules, providers and deployers of these systems already face transparency obligations if they generate or manipulate content to look like real people. Deepfake outputs should be clearly labelled as AI-generated, especially when they are published to inform the public.EU transparency guidance

The Denmark model goes a step further by saying: beyond labelling, is the underlying use of your likeness even allowed?

For example, in a Denmark-style regime:

  • A business using an AI avatar that looks and sounds like a real employee would likely need explicit consent and possibly a written licence.
  • An app that lets other users generate explicit content with your face from a stolen selfie would be clearly illegal to distribute, not just “problematic” or “against platform policy.”
  • A creator could license a digital double of themselves, then enforce that deal the way musicians enforce song copyrights.

You still need to read the terms of service, but the law starts to backstop your rights rather than leaving everything to platform policy.

The Opportunities: Not Just Protection, But Ownership

The conversation about deepfakes is often framed around harm – and that harm is very real. But the Denmark model hints at an upside too: monetizable ownership of your digital self.

If your likeness is treated more like a copyrightable asset, you could imagine:

  • Personal “likeness licences”
    You grant a game studio the right to use your face as an NPC for a fee, or let a virtual production company cast your digital double as background characters.

  • Creator-centric digital doubles
    Influencers and streamers could license different “versions” of themselves for branded content, international dubbing, or 24/7 AI-powered fan interactions – with contractual guardrails on what those doubles can and cannot say.

  • Workplace avatars with boundaries
    You agree that your employer can use a cloned voice to handle some customer calls, but only for a defined time and domain, and only as long as you’re employed, with compensation baked in.

All of this is much easier to negotiate if the baseline law says: your likeness is yours by default, and others need your permission to use it.

The Risks and Open Questions

The Denmark model isn’t a magic wand. There are real challenges and tradeoffs:

  • Freedom of expression and satire – Laws must carve out space for parody, art, and political commentary. The EU AI Act already notes that deepfake labelling rules shouldn’t be interpreted as restricting freedom of expression when content is clearly artistic or satirical, as long as rights of third parties are safeguarded.AI Act recital on deepfakes

  • Enforcement across borders – Denmark can pass its own law, but deepfake-generating apps can be run from anywhere. This is where EU-wide rules and international coordination matter.

  • Who counts as “you”? – What about stylized or partially anonymized avatars? What if an AI generates a face that is not you but looks a lot like you? Legal scholars are already debating how broadly “likeness” should be defined in AI law.Analysis of deepfake definitions

  • Power imbalances – Just because rights exist on paper doesn’t mean everyone can easily use them. Without accessible takedown systems and support, only those with lawyers and time might fully benefit.

Still, the Denmark model signals a shift: from “we will moderate if things go wrong” to “we recognize you as the rights-holder of your digital self from the start.”

What You Can Do Now to Protect (and Own) Your Digital Likeness

You don’t have to wait for your country to become “the next Denmark” to get more control. You can start with three practical steps:

  1. Audit where your face and voice live online.
    Search your name and images regularly. Check which apps you have allowed to access your photos, camera, and microphone. Revoke permissions you don’t need and delete accounts you no longer use, especially in sketchy AI “photo enhancer” or cloning apps.

  2. Choose AI tools with clear transparency and rights policies.
    Before you upload a selfie, sample, or voice recording to any AI service – whether it’s a mainstream model like ChatGPT and Gemini or a niche avatar app – read how they handle training data, likeness rights, and takedown requests. Prefer tools that explicitly give you control over how your likeness is used and allow you to opt out of model training.

  3. Support – and demand – Denmark-style rights where you live.
    As lawmakers everywhere scramble to respond to deepfakes, you can push for policies that align with the Denmark model: clear ownership of your likeness, explicit consent for AI imitations, and strong takedown mechanisms. When public consultations or platform policy feedback forms open up, mention these principles explicitly.

You may not care about being famous, but in the AI era, your digital self is already out there, training models and populating feeds. The real question is whether you simply exist in those systems – or you actually own your piece of them.