You are sitting quietly on the sofa when suddenly you see a video in which you, or someone who looks exactly like you, says something you never said. Or you hear your voice in an advert you never participated in. These are deepfakes: images, sound or text that have been created or manipulated using artificial intelligence. They look real, but they are fake. And it is often almost impossible to judge what is real and what is not.
Deepfakes are not just videos or photos: they also include voice clones and even AI-generated texts. According to the European AI Act, this refers to any AI-produced image, audio or video material that is so realistic that it can be mistaken for the real thing. Think of the viral imitation of Tom Cruise on TikTok, deepfake videos of Putin or Zelensky in the war, or so-called “nudify” apps that create nude images without consent. Currently, the vast majority of harmful deepfakes consist of non-consensual pornography – with far-reaching consequences for victims.
In Denmark, there is already a proposal to give people copyright on their own face and voice. And in the Netherlands, too, work is now underway on similar protection. But via a different route: the neighbouring right.
A growing legal and social problem
Deepfakes can be used in a humorous or creative way, for example in advertisements or films, provided permission has been obtained. However, they are also a source of deception, fraud and reputational damage. In Hong Kong, a financial employee was deceived by a deepfake video call with ‘colleagues’, causing him to transfer 25 million dollars to fraudsters. Deepfakes are now also being used as evidence in court cases, which increases the importance of verification by experts.
Until now, protection in the Netherlands has been fragmented:
- Portrait rights only apply to recognisable images, and not to voices.
- The GDPR protects personal data, but not the voice or appearance of deceased persons.
- Only sexual deepfakes are explicitly protected by criminal law.
- Other cases must be dealt with through civil law, often involving lengthy proceedings.
A new bill: control over your own deepfake
To bridge this gap, Member of Parliament Rosemarijn Dral has submitted a private member’s bill in October 2025. This proposal (not yet law) introduces a new neighbouring right for natural persons. This gives everyone the exclusive right to determine whether and how their voice or appearance may be used in a deepfake. This means that it is not permitted without consent.
The proposal adds an article to the Neighbouring Rights Act and gives people three important powers:
- allow or prohibit the creation of deepfakes;
- determine whether it may be made public or distributed;
- set conditions for commercial use.
Furthermore, the right remains valid for seventy years after death, so that surviving relatives can also protect the reputation and legacy of a loved one. This is new: there were no specific regulations for “post-mortem deepfakes” in the Netherlands before.
Why a neighbouring right?
This approach is in line with the European tradition of regulating new technological developments within existing intellectual property law. Previously, for example, producers and press publishers were already granted their own neighbouring rights. Now, this would be extended to natural persons themselves.
The advantage of this is that the right is internationally recognisable and aligns with existing performance rights. For artists, actors, and singers, this provides a familiar legal framework for maintaining control over their voice and appearance, including in commercial collaborations such as hologram performances or the use of AI voices in advertisements.
Important: this right is non-transferable. It is personal and cannot be sold. You can grant permission or a licence, for example for a film or music recording, but you always retain control.
How does this relate to the AI Act?
The European AI Act will require providers of AI systems to make deepfakes recognisable from August 2026 onwards, for example through watermarks or metadata. Users who publish AI content must also state that it is artificially generated content.
But that AI regulation is not about consent or control. It only regulates transparency. The Dutch proposal fills that gap: it determines who can actually decide on the use of your voice or face.
What can you do if you are the victim of a deepfake?
Anyone who encounters a deepfake can take action themselves:
- Record evidence (screenshots or downloads).
- Report the video to the platform where it is posted.
- Consider reporting it to the police.
- Seek legal advice, for example from a lawyer, to demand removal or a ban.
Finally: towards digital empowerment
The deepfake bill is not yet in force, but it marks an important moment. As technology becomes increasingly adept at copying people, the law is attempting to restore the boundaries of identity and autonomy.
Your face and voice are not fair game for AI; they belong to you. And if someone misuses them, the law should protect you.
Need help removing a deepfake of you from the internet? Contact lawyer Julia van Leeuwen.
