Denmark's proposed copyright rules on deepfakes

Denmark's proposed copyright rules on deepfakes

Denmark has notified a draft act amending its Copyright Act to address realistic digitally generated imitations, commonly known as 'deepfakes'. The draft was submitted via TRIS (the Technical Regulation Information System), a database used by EU Member States to submit certain draft regulations under Directive (EU) 2015/1535. The Danish notification is registered as 2025/0654/DK. According to the public TRIS record, the European Commission issued comments on 3 February 2026.

A deepfake is generally understood to be AI-generated or AI-altered audio, image, or video content that realistically imitates a person. Such content may be harmless in some contexts, for example for satirical or entertainment purposes. However, it can also be used to mislead the public, damage reputations, or imitate a person’s voice or appearance without consent.

The Danish proposal would introduce two forms of protection to the Copyright Act. The first would protect individuals against the publication of realistic digital imitations of their personal physical characteristics, such as their appearance or voice, without their consent. Furthermore, it would safeguard performers and artists against the unauthorised creation of realistic digital reproductions of their performances or artistic presentations. As set out in the TRIS notification, such unauthorised publication would be illegal under Danish law, and the content could be reported to online sharing platforms as illegal content.

In addition, the draft contains safeguards for certain forms of expression. For instance, the proposed protection for personal characteristics would not apply to imitations that constitute mainly caricature, satire, parody, pastiche, criticism of power or social criticism, unless the imitation constitutes misinformation that could seriously endanger the rights or essential interests of others. The draft text states that the proposed Act is expected to come into force on 1 July 2026.

The proposal is legally significant because it uses copyright law to address issues linked to identity, privacy, reputation and public trust. The European Parliamentary Research Service has described the Danish initiative as a novel approach that would provide personality-related protection through copyright law. While there is broad interest in addressing harmful deepfakes, the most appropriate legal framework and the need for new rights are still being debated.

How this fits within the EU framework

The Danish proposal should be read alongside the existing EU rules. The AI Act already includes transparency obligations for certain AI-generated or AI-manipulated content, including deepfakes. The Commission is also working on a Code of Practice on marking and labelling of AI-generated content to support compliance with these obligations.

Other EU instruments may also be relevant. The Digital Services Act sets out the rules for notifying online platforms of illegal content. The Copyright in the Digital Single Market Directive includes specific provisions for online content-sharing service providers about copyright-protected content. Therefore, Denmark’s proposal sits at the intersection of copyright, personality protection, platform regulation, and AI transparency.

Other approaches in the EU and internationally

Other EU Member States have adopted different legal approaches to address the risks associated with deepfakes. For instance, France has amended its Criminal Code to address the dissemination of algorithmically generated visual or audio content representing a person’s image or words without their consent, where the artificial nature of the content is not obvious or expressly mentioned. France has also introduced a specific offence for sexual deepfakes. Italy’s 2025 AI law introduced criminal provisions concerning the unlawful dissemination of AI-generated or AI-altered content. These examples show that Member States are not using a single legal model.

Outside of the EU, the United States has adopted the TAKE IT DOWN Act, which addresses non-consensual intimate visual content, including that generated by AI. The Act requires covered platforms to implement a notification process and remove such content within 48 hours of receiving a valid notification.

China has adopted a regulatory model that places a strong emphasis on content governance and labelling. Its rules on deep synthesis services, along with subsequent measures regarding the  labelling AI-generated synthetic content, require providers to identify such content and minimise the risk of users confusing it with authentic content.

South Korea has focused on criminal law and specific risk areas. Recent measures have addressed sexually explicit deepfakes, and election-related rules have been used to tackle deceptive AI-generated content in politics. Furthermore, South Korea has announced plans to require the labelling of AI-generated advertising to combat misleading promotional content. 

Japan has taken a more framework-based approach. Its first AI-specific law focuses on promoting the development and use of AI while allowing government action in response to serious incidents. Japanese authorities have also begun examining civil liability for unauthorised uses of a person’s likeness or voice in AI-generated content, including possible links with portrait rights and publicity rights.

What this means for businesses 

The main practical point for SMEs is that deepfake regulation is developing across several areas of law. Depending on the circumstances, the relevant rules may concern copyright, performers’ rights, personality rights, data protection, criminal law, platform responsibilities, or AI transparency obligations. Businesses using generative AI should therefore ensure that they have the necessary consents to use individuals' voices, images or performances, and that AI-generated or AI-altered content is labelled as such.

Denmark’s proposal does not resolve the broader EU debate. However, it shows that lawmakers are seeking clearer tools with which to address realistic digital imitations. The next key question is how national initiatives can be aligned with existing EU rules while preserving legitimate uses, such as parody, satire, criticism and artistic expression.