
The Netherlands has no dedicated deepfake law, but your face is protected: through portrait rights, the GDPR and the AI Act. Creating or using an AI version of a real person means dealing with a patchwork of rules. Existing rules reach further than many people think, though, and new legislation is on the way. An overview for talent, brands and managers.
What already applies in the Netherlands
No Dutch law bans creating a deepfake as such. According to the Dutch government, using AI to imitate someone without consent will in practice almost always be criminal or unlawful. Protection comes from three directions:
- Portrait rights (Article 21 of the Dutch Copyright Act). For a portrait not made on commission, which covers most deepfakes, you can block publication if you have a reasonable interest. Commercial use without consent is a classic example.
- The GDPR. Your face and voice are personal data. Creating and distributing a deepfake without a valid legal basis breaches privacy law, and you can file a complaint with the Dutch Data Protection Authority.
- Criminal and tort law. Sexual deepfakes made without consent are a criminal offence. Misleading or harmful deepfakes can be a tort.
Portrait rights have clear gaps. They don't protect against voice clones, it is unclear how they apply after death and every case depends on balancing interests after the fact.
The proposal: your own right over your deepfake
To close those gaps, Dutch MP Rosemarijn Dral (VVD) presented a private member's bill on 30 October 2025 for a 'neighbouring right' over deepfakes of persons. Everyone would get the exclusive right to authorise the creation and distribution of a deepfake of themselves, voice included, lasting until seventy years after death.
The public consultation closed on 31 December 2025. Dral did not return to parliament after the elections and responses from the field were largely critical. Professor Bernt Hugenholtz called it the right idea in the wrong legal framework. Whether and when the bill reaches parliament is unclear. Denmark is pursuing a similar route through copyright law.
Europe: labelling and a new ban
At EU level the AI Act is the main rule. Since 2 August 2026, anyone publishing a deepfake must disclose that the content is artificial, as explained in our piece on Article 50 and labelling. Through the Digital Omnibus the EU also added a new ban: from 2 December 2026, AI systems that generate child sexual abuse material or non-consensual intimate imagery are prohibited.
Note: the AI Act is mainly about transparency. It does not say who must consent to the use of your face. For that you still rely on portrait rights, the GDPR and your contract.
United States: the NO FAKES Act stalls
The US NO FAKES Act would create a federal right over your voice and likeness in AI replicas, including a takedown system. The right could not be assigned during your lifetime, only licensed. On 18 June 2026 the Senate Judiciary Committee advanced it unanimously. An attempt on 30 September to pass it through the full Senate by unanimous consent was blocked by an objection from Senator Ted Cruz. The bill is on hold until there is agreement on free speech protections.
States are moving faster. Tennessee protects voice and likeness through the ELVIS Act. In New York, since June 2025 brands and model agencies need written consent before creating a model's digital replica.
What this means in practice
While the law lags behind, the contract is your main protection. Anyone working with an AI likeness should at least agree on:
- explicit consent per use, separate from ordinary image rights;
- which layers may be used: image, voice, movement;
- term and deletion of the data afterwards;
- a ban on training other AI models with the data;
- who handles AI disclosure under the AI Act;
- a misuse clause with fast takedown and penalties.
Why a licence beats an assignment here is explained in license or sell your likeness.
Where Adapt Legal fits
Adapt Legal builds exactly this layer: campaign contracts, a log of every consent and enforcement when misuse happens. Talent and brands don't have to wait for lawmakers.
Are you talent and want control over your image? Read how a digital twin works for you.
This article is general information, not legal advice.
Frequently asked questions
Is creating a deepfake illegal in the Netherlands?
Creating one is not specifically banned. According to the Dutch government, imitating someone with AI without consent is in practice almost always criminal or unlawful, through criminal law, the GDPR or portrait rights. Sexual deepfakes without consent are a criminal offence.
Do Dutch portrait rights cover my voice?
No. Portrait rights cover images. Against a voice clone you need to rely on the GDPR, tort law or solid contract terms.
Sources
- Dutch government answer to parliamentary questions on deepfakes, 15 April 2026 (Officiële Bekendmakingen)
- Explanatory memorandum, Dutch bill on deepfakes of persons (internetconsultatie.nl)
- Een naburig deepfake-recht. Echt? (P.B. Hugenholtz, NJB 2026, via IViR)
- The AI Act implementation timeline: what changes under the AI Omnibus (Future of Privacy Forum)
- NO FAKES Act of 2026, S.4591 as reported (GovInfo)
- NO FAKES Act "dead for now" as Cruz objects to unanimous consent request (Daily Journal)
- New York Fashion Workers Act: consent requirements for digital replicas (Benesch)