Skip to content

Experts insights

Do we really need to copyright our own faces?

TL;DR

  • Denmark wants to fight AI deepfakes by giving every citizen copyright over their face, voice and body.
  • The approach works better than it should: it lowers the burden of proof for victims and plugs into the EU's takedown mechanism.
  • It isn't really copyright, though, not even by Denmark's own admission, and the European Commission agrees.
  • The durable fix will likely combine five things: a standalone personality right, watermarking and provenance standards, EU-wide harmonization, stronger enforcement of existing law, and platform accountability that doesn't depend on proving infringement.

Denmark appears to have found a solution to AI-generated deepfakes by granting every citizen copyright over their face, voice and body.

This measure may seem extreme when you consider that a person's appearance can hardly be treated as a form of intellectual property.

However, generative AI is irreversibly undermining the concept of a trustworthy image, so it is worth exploring this issue further.

The Danish Proposal

The plan is to amend Danish copyright law so that individuals gain control over realistic digital replicas of their personal characteristics, with parallel protection for artists and performers against synthetic reproductions of their performances. Consent is the key: publishing this type of content without the individual's consent gives them the right to demand its removal and, in certain cases, compensation without having to prove reputational harm. The protection would last for fifty years after death and would be enforced through the mechanisms of the European Digital Services Act rather than criminal law. Parody, satire, caricature and pastiche would remain excluded, at least in principle.

This approach is elegant: copyright already has a robust enforcement infrastructure in place, including notice-and-takedown procedures, platform liability and established remedies. Furthermore, the public understands what constitutes an infringement of rights. Borrowing this mechanism to protect identity means that victims do not have to build a case from scratch each time but can instead follow a familiar procedure by requesting the removal of abusive content.

However, this elegance is partly an illusion, as the Danish government is aware.

The Legal Dilemma

According to the European Parliament’s Research Service, the proposed rules do not constitute copyright in the strict sense of the term. In the words of the Court of Justice of the European Union, copyright protects original works that reflect the author’s own intellectual creation. A voice or an image don't fall under this definition. Moreover, in the case of photographs, ownership normally belongs to the creator, that is, the photographer, not the person portrayed. Danish authorities admit that they have borrowed the 'toolkit' of copyright law because its tools seem to offer an easy solution to the problem.

Nevertheless, copyright laws are not suited to protecting personal rights because:

  • Copyright protects creative expression, not the mere fact of existence. This is why a face has never qualified as a work.
  • It rewards the creator; therefore, attributing it to the subject of the image overturns the usual logic of ownership.
  • It can be bought, sold and licensed, which risks turning faces and voices into tradable commodities — the exact opposite of what an inalienable personality right should do.

These are just some of the reasons why several Member States have been hesitant, arguing that the issue of deepfakes extends far beyond the scope of copyright. Meanwhile, in February 2026, the European Commission, noting a conceptual misalignment with the purpose of intellectual property, urged Denmark to achieve this objective through other means.

Further concerns have also been raised about the implications of applying copyright to the human body, as this would effectively redefine it as a property that can be bought and sold. However, the issues at stake, such as honour and dignity, fall under personality rights rather than property rights.

Finding the Right Solution

Existing regulatory frameworks share a common limitation: they were designed for a world in which impersonating a specific person was difficult, expensive and uncommon. That world no longer exists. It is estimated that there were around eight million deepfake files in 2025, compared to about half a million in 2023. While the nature of the harm remains unchanged, its scale has grown exponentially. Asking the victim to prove reputational harm, commercial use or criminal intent is too much to ask of someone who has found their face spliced into a video by an anonymous account and replicated across a dozen platforms. The friction that once protected privacy now protects those who produce the fakes.

Despite its doctrinal imperfections, what Denmark offers is a lower threshold and a faster path. By linking harm to consent rather than proof of prejudice, it enables individuals to object without first having to demonstrate that they have been harmed. By integrating with the DSA’s notice-and-action mechanism, it provides a mechanism for enforcing that “no”. There is also a more subtle benefit: a clear consent requirement has a preventive effect even before removal is necessary because it changes what platforms and content creators believe they can get away with.

However, it is evident that the solution must be found elsewhere. The most promising approaches include:

  • An autonomous right of personality, designed from the ground up to protect identity and dignity, rather than grafted onto a framework created for creative works.
  • Infrastructure-level measures, such as watermarking and content credentials for synthetic content, along with the strengthening of the labelling requirements already provided for in the AI Act.
  • Harmonization at the EU level, transforming the Danish initiative into a category of illegal content recognized throughout Europe, thereby overcoming the limitations of geo-blocking.
  • More decisive enforcement of existing tools (GDPR, criminal law for the most serious cases, unfair competition), rather than the proliferation of new rights.
  • Holding platforms accountable in a way that does not depend on proving a violation of a specific right, but treats rapid removal as a digital product safety obligation.

Conclusions

Do we really need to copyright our own faces? This remains an uncomfortable question and is clearly not a viable solution to the problem of AI-generated deepfakes. However, the fact that a government has felt compelled to address this issue with such an extreme measure suggests that the current approach is no longer an option.

Denmark deserves credit for being the first to spark this debate. Ireland has suggested that it may follow, and the European Parliament is moving toward its own work on copyright and generative AI.

Finding the right solution quickly is now essential and will probably involve a combination of factors to guarantee citizens’ safety while sharing responsibility with the platforms.

Clearer decisions. Better budgeting. Fewer losses.

From "how many attacks?" to "how much damage?"
The clarity needed for sharper strategy.