Why Denmark’s proposed deepfake law is still on hold
Imagine discovering an advertisement in which a digital version of you promotes a product you have never used. The face and voice are yours, but you never took part.
In July 2025, Denmark proposed new rules to give people greater control over realistic AI imitations, often called deepfakes. Section 73a would protect anyone against publication of an AI imitation of their appearance, voice or movements without consent. Section 65a would give performers and artists similar protection when AI imitates a performance. Neither provision is in force.
A problem that is already here
DR recently reported that advertisements on Facebook and Instagram used images of real people, including children, in sexualised AI-generated videos. Of nine suspicious ads reviewed with Red Barnet’s AnmeldDet service, four were reported to police as suspected child sexual abuse material. The reporting prompted calls for the Justice Minister to explain what Danish authorities can do and whether companies connected to the advertising can be held criminally liable.
That example concerns the most serious end of the spectrum. The same basic issue can affect a public figure, an employee or a private person: who authorised the use of the face, voice or performance?
Why the Commission became involved
Denmark had to notify the European Commission before adopting rules that could affect digital services across the EU. The notification started a temporary standstill period, giving the Commission and other EU countries time to check whether the draft conflicted with EU law or created barriers within the internal market. Denmark submitted the proposal on 31 October 2025.
The Commission issued comments on 2 February 2026. They were not a veto, and the standstill ended the next day. The Commission nevertheless questioned whether personal characteristics belonged in copyright law, whether the performer right overlapped with existing EU rights, and whether the proposal fitted with EU rules for platforms and AI-generated content.
The draft did not reach the Folketing before the general election on 24 March. The formal EU standstill is over, but the Danish legislative project is still standing still.
Why the delay matters
Danish law does not leave victims without protection. Depending on the circumstances, identity misuse, sexual deepfakes or defamation may already be covered by criminal law. Marketing, personality and data-protection rules may also apply.
The proposal was designed to provide a clearer civil route. The affected person or performer could invoke a specific right and ask for the material to be removed. The draft was expressly intended to support removal from social media under the EU Digital Services Act, which regulates how platforms respond to illegal content. Courts could order removal, and performers could also claim payment and damages.
This could make enforcement more direct and, in some cases, faster than first establishing criminal liability. It would not guarantee a quick outcome in every case. The main practical benefit would be a clearer legal basis for demanding removal.
What happens next
The new Government has promised legislation on ‘AI protection of personal characteristics’, but no replacement draft or timetable has been published. Sections 65a and 73a therefore remain proposals, and the final rules may look different.
Until then, anyone publishing AI-generated images, voices or video should remember that labelling content as AI-generated does not itself provide permission to use another person’s identity.
Get in touch
aera advises on AI-generated content, digital identity and performance rights. Please contact us to discuss the legal implications of deepfakes.
Sources: DR investigation; notified Danish draft; Commission comments C(2026) 756; June 2026 Government platform; and EU AI Act.
