How long does it take before a deepfake of me is offline?

There is no fixed deadline. The Digital Services Act requires a platform to handle a notice about a deepfake in a timely and diligent manner and to tell the person who reported it what it has decided, but it sets no number of hours or days. If the platform and the uploader do not cooperate, interim injunction proceedings follow. The hearing before the voorzieningenrechter, the judge who sits in urgent cases, is usually held within two to six weeks, and judgment generally follows within two weeks after that. In very urgent cases the court can decide sooner.

What does it cost to have a deepfake of me removed?

A notice to the platform and the Offlimits helpline cost nothing. If interim injunction proceedings are needed, expect €5,000 to €20,000 excluding VAT at Liaise for the whole procedure, from intake to hearing. Court fees come on top, and they are lower for a private individual than for a business. If you win, the court will usually order the other party to pay your legal costs. That award follows fixed scales and rarely covers everything you have spent.

Can I claim damages from whoever made a deepfake of me?

Yes. Anyone who makes or spreads a deepfake of you generally acts unlawfully under Dutch tort law (article 6:162 of the Dutch Civil Code) and processes your personal data without a legal basis, for which article 82 of the GDPR gives a separate right to compensation. Material loss, such as a lost commission, is compensated if you can prove it. For non-material harm you have to show that you have been affected as a person. With a serious infringement, such as a sexual deepfake, the court accepts that more readily, but the amounts Dutch courts award are modest. You also need to know who made or posted the deepfake.

May someone make a deepfake of me as a joke or satire?

Satire is covered by freedom of expression (article 10 ECHR) and is given room for that reason. The court weighs that freedom against your right to privacy and your reputation. A deepfake that is clearly recognisable as satire and concerns your public role is more likely to stand than an image that looks real or shows you in a sexual situation. Since 2 August 2026 the AI Act also requires a disclosure that the image is artificial for evidently satirical work, although that disclosure must not hamper the enjoyment of the work.

Having a deepfake or cloned voice of yourself removed

A video in which you say something you never said. A nude image someone made with AI from an ordinary holiday photo. Getting a deepfake of yourself removed usually works, provided you know whom to approach and on what legal ground. The same goes for a cloned voice in an advert you never set foot in a studio for.

Liaise acts for people who have been imitated with AI, from private individuals to artists, presenters and voice actors. We have the material taken down, find out who is behind it and bring interim injunction proceedings where needed.

Is a deepfake of you going around right now? Call 020 675 88 21.

Record what is circulating before you have a deepfake removed

Once a platform takes a deepfake down, your evidence goes with it. So start by taking screenshots of the image or video, with the URL, the date and the name of the account that posted it. Keep comments and reposts too, because they show how far the material has spread. That counts when a court later has to decide how urgent an injunction is.

Don’t forward a sexual image to ask for help, not even to friends. That widens its reach. The helpline of Offlimits, the Dutch expertise centre on online abuse, gives free and anonymous help with unwanted images online.

Having a deepfake removed by the platform

The quickest route is often through the platform itself. You don’t need to know who made the deepfake for that. Under the Digital Services Act (DSA) every platform must have a notice system. If your notice points to the deepfake precisely and explains why it is unlawful, it counts as knowledge on the platform’s part (article 16(3) DSA). From then on the platform can only rely on the rule that it is not liable for what its users post if it removes the deepfake promptly (article 6 DSA).

The platform has to handle your notice in a timely and diligent manner. You will hear what it decides. A notice that only says a video is fake often goes nowhere. Platforms take a deepfake down sooner when the notice names the legal ground and explains how the image harms you.

If the deepfake stays up on a site that does not respond, you can ask Google to remove the page from the search results for your name. You do that with your right to erasure under article 17 of the GDPR.

The legal grounds for acting against a deepfake of yourself

Your face and your voice are personal data as soon as they are recognisable. If someone processes them without a legal basis, that breaches the GDPR. You can then demand that the image be erased (article 17 GDPR) and claim compensation (article 82 GDPR). In many deepfake cases the GDPR is the strongest ground, because the maker almost never has a legal basis.

If you recognise yourself in a deepfake, it is also a portrait of you. Under Dutch image rights it may not be published if you have a reasonable interest in opposing that (article 21 of the Dutch Copyright Act). With an image that shows you doing something you never did, that interest is obvious. A portrait does not have to be real. According to the Dutch Supreme Court, an image of a lookalike can be your portrait too (22 April 2022, ECLI:NL:HR:2022:621). That reasoning applies just as well to a face that AI has generated.

Tort (article 6:162 of the Dutch Civil Code) covers what the GDPR and image rights leave open. An example is a fake post with a generated quote that damages your honour and reputation without any image of you in it. You can read more about the balancing exercise the court then carries out under unlawful publication and image rights.

What the court decided against Grok in March 2026

On 26 March 2026 the voorzieningenrechter in Amsterdam, the judge who sits in urgent cases, showed how far the GDPR reaches. Offlimits sought an injunction against the feature of the chatbot Grok that let users undress people in photos. The court prohibited xAI from making or distributing sexual images of identifiable people living in the Netherlands without their consent, subject to a penalty payment of €100,000 a day up to a maximum of €10 million. X may not offer Grok for as long as Grok fails to comply with that order (ECLI:NL:RBAMS:2026:3106).

The ground was the GDPR. According to the court, Grok processes personal data the moment it creates an image with someone’s face without checking whether that person consented. The court left open whether xAI is itself liable for the images. For anyone who has been imitated the judgment still matters, because you don’t have to wait until you know who gave the instruction. The provider of the AI system can be held to account as well.

Acting against a cloned voice

A voice is not an image. That is why a voice falls outside image rights. A recording of your voice is personal data, though, and so is a clone that sounds like you. If someone uses that clone to sell something, that is quickly unlawful. Nobody has to accept their voice being linked to someone else’s product without consent.

In April 2026 the Belgian voice actor Nicolas Cherlet heard himself in an online advert for an estate agent. He had never recorded that text. The production company had cloned his voice from recordings he had made for it earlier. The Belgian association of voice actors combined his story with those of colleagues in a complaint to the Belgian data protection authority.

If you work as a voice actor, artist or presenter, set out in your contract whether your voice may be used for AI. Our article on the use of voices by AI explains how the law protects your voice.

When a deepfake is a criminal offence

Making a sexual deepfake is a crime. Article 254ba of the Dutch Criminal Code makes it an offence to intentionally and unlawfully produce a visual depiction of a sexual nature of a person, with a maximum of one year’s imprisonment. Publishing it carries up to two years, where the person spreading it knows that the image may harm the person shown.

Other deepfakes and a cloned voice are not offences as such. What is done with them can be, such as fraud using the imitated voice of a family member, or defamation. Reporting it to the police and bringing civil proceedings do not rule each other out. Criminal law will not take an image offline for you, though. The Public Prosecution Service also decides for itself whether to prosecute.

If you don’t know who made the deepfake

Many deepfakes come from anonymous accounts, whose details a platform will not simply hand over. In 2005 the Dutch Supreme Court set out when a provider does have to disclose them: where it is plausible that the publication is unlawful and harmful, you have a genuine interest in the details, there is no less intrusive way to obtain them and your interest outweighs that of the provider and the uploader (Lycos/Pessers, ECLI:NL:HR:2005:AU4019).

If the platform refuses, you ask the court to order disclosure. That can be done in the same interim injunction proceedings as the removal.

The AI Act and deepfakes: labels and a ban

Since 2 August 2026 a deepfake must be recognisable as artificial (article 50(4) of the AI Act). A label does not make the deepfake lawful. You can still have a video removed that is neatly marked as AI, if it infringes your image rights or your privacy.

From 2 December 2026 the AI Act also prohibits AI systems that produce sexual or intimate images of identifiable people without their consent. That ban was added to article 5 by Regulation (EU) 2026/1744. A provider that places such a system on the market without adequate safeguards risks a fine of up to €35 million or 7% of worldwide annual turnover.

The Netherlands has no separate right in your face and voice of the kind Denmark wants to introduce with a bill tabled on 8 October 2026. The Dutch private member’s bill for a neighbouring right in deepfakes went out for consultation in October 2025, drew a lot of criticism and has not been taken further since.

When you have to label AI content →

Deepfake lawyer in the Netherlands

Liaise acts for people who have been imitated with AI and for creators who work with digital doubles themselves. If a deepfake of you is going around, we draft the notice to the platform, send a cease-and-desist letter to whoever posted it and bring interim injunction proceedings where needed, including against the provider of the AI system. If you work with a digital double of an actor yourself, read what applies to deepfakes in film.

Is a deepfake or a cloned voice of you going around? Get in contact or call 020 675 88 21. You will have an answer within one working day.

Practice areas

Merel Teunissen

Lawyer

Merel advises and litigates in the fields of media law, film law, copyright, privacy law, contract law and intellectual property law. She acts for established and emerging media personalities, producers in film, television and podcasts, victims of unlawful publications and creative entrepreneurs.

Charissa Koster

Lawyer

Charissa advises and litigates in the fields of media law, intellectual property, IT and AI law, privacy, e-commerce and corporate law. Press-related and journalistic matters, copyright, freedom of expression versus privacy and cross-border issues often play a central role in her work.