Denmark, Deepfakes, and a New Crisis in the Digital Sphere

A few months ago, Danish newspaper Politiken, along with the Canadian public broadcasting corporation, CBC, and Dutch Bellingcat revealed the real identity of AI porn magnate “MrDeepFakes”. The man, whose real name is David Do, has made thousands of deepfake porn videos seemingly starring several famous women such as the queen of Denmark and climate activist Greta Thunberg.

Due to the rapid evolution of AI technology, creating these kinds of custom videos is an easy thing to do for the somewhat tech savvy. Around the MrDeepFake site a community arose between makers and buyers of this content, as well as people who discussed ways to improve themselves as virtual pornographers. The site has been taken down after the revelation of David Do’s identity, but a simple Google search reveals new mirror sites promising to rebuild the old community, and other sites where people can watch similar content. Like a digital hydra, one head was cut off and many more appeared. As is often the case with these technical advances, many probably remember the legislative issues that arose with digital piracy in the 00’s.

The deepfake phenomenon, much like traditional revenge porn, disproportionately targets women and public figures. Now, lawmakers argue that granting individuals copyright over their likeness could also serve as a tool against gender-based digital violence.In the forefront of the battle against deepfakes is Denmark. The country’s parliament is currently preparing to vote on a new law broadly limiting the ways that people may produce deepfakes of others in the future. While the EU already has a law, passed last year, requiring makers and sharers of deepfake material to disclose its authenticity using watermarks, the Danish bill promises to take things even further. If this law is adopted, Danish citizens will have copyright over their own “likeness”. This goes for the face as well as body and voice. Individuals who come across deepfake content of themselves can demand a takedown in accordance with this self-copyright law. The bill tackles the question of what it means, culturally, to ‘own’ your face or your voice in a digital economy. In our time, digital identity can be copied, sold, or simulated. Denmark’s bill touches not only law, but also the question of how we define authenticity in an era of infinite reproduction.

On another note, France was among the first European countries to regulate AI-generated content. Its new law for the digital spaces, adopted in 2024, requires that AI-produced images, audio, and video be clearly labeled as such. It also prohibits distributing AI-generated depictions of individuals without their consent, with penalties for platforms that fail to act in accordance with the law. However, unlike Denmark’s proposal, the French rules focus on platform responsibility and transparency rather than granting individuals ownership rights over their likeness.

Yet as technology advances, will there come a pan-EU solution?

The European Union’s own AI act was passed in 2024. Parts of it have already come into effect, and other parts will come into effect in August of next year and then more further down the line. It’s a starting point similar to the French legislation that requires AI generated content to be branded as such. The purpose is to set a European standard, and to introduce definitions and risk levels to aid lawmakers in the future. The EU law is not to position itself in the forefront of these issues.

Perhaps that’s the position that Denmark will take? 

The same week that the story of David Do was published by the aforementioned news outlets, the Romanian MEP Dan-?tefan Motreanu asked the European Commission about what concrete measures they intend to take to address risks posed by deepfake and AI-generated content. In the question he cited the Danish bill as a good example, and a handful of countries have discussed similar legislation for themselves. In the Netherlands, several parties have introduced bills to grant citizens copyright over their own bodies and voices. Such legislation is supported by a majority of the Dutch house of representatives, but there are still a lot of things waiting to be decided, and new legislation will likely not be introduced in the Netherlands until 2026, which would be considered fast.

However, recently a heavy name in the EU, Mario Draghi, lifted concerns about the implementation of several parts of the AI act. In a speech delivered precisely a year after his famous Draghi-report, he even went as far as to call for a pause of the law’s implementation. This to ensure European competitiveness in the AI market. He stated:“Implementation of this stage should be paused until we better understand the drawbacks.”

The commission has since claimed that there will be no pause in implementing the AI act, opposing Draghi’s suggestion. The European Commission’s statement is backed by the over fifty organizations, including CDT and BEUC, who’ve previously raised concerns about slowing down the regulation on these integrity compromising technologies.

Europe’s struggle to govern AI is not only a technical one. It’s a cultural question about who we are in the digital age, and also who controls the stories told about our faces, our voices, and our bodies. If Denmark succeeds in setting a new precedent, it could redefine not only copyright, but the very notion of digital personhood across Europe.