A media company in Copenhagen launches a content platform powered by a large language model. Its editorial team curates prompts, selects outputs, and edits the results. Months later, a competitor reproduces the platform's AI-generated articles verbatim. The company's legal team reaches for copyright law – and finds the ground far less solid than expected. Denmark's intellectual property legislation was drafted for human authors. It is now being stress-tested by outputs that no human authored in any traditional sense.
Danish intellectual property legislation conditions copyright protection on an original creative work produced by a human author. AI-generated outputs do not qualify automatically. Whether any protection attaches depends on the degree of human creative input embedded in the generation process – a factual question that Danish courts and regulators are only beginning to address systematically.
This analysis examines the doctrinal foundations of Danish copyright law as applied to AI-generated works, surveys the gap between current statute and emerging practice. Considers the EU-level regulatory overlay. Additionally, draws out the strategic implications for international businesses operating in or through Denmark.
Doctrinal foundations: what Danish copyright law requires
Danish intellectual property legislation – rooted in the ophavsretsloven (Danish Copyright Act) – protects literary, artistic, and scientific works that bear the stamp of individual creative expression. The originality threshold is not demanding by international standards. A short poem or an idiosyncratic photograph can qualify. What the legislation does not do is extend protection to outputs that lack a human author entirely.
This human-authorship requirement is not unique to Denmark. It reflects a principle shared across the Nordic countries and broadly consistent with EU harmonisation directives on copyright. The Sø- og Handelsretten (Danish Maritime and Commercial Court), which handles a significant share of intellectual property disputes in Denmark, has not yet issued a definitive ruling on AI-generated outputs as such. However, the court's existing body of case law on computer-assisted works and software-generated materials points firmly toward a human-creative-contribution test.
That test asks a specific question: did an identifiable human being make creative choices that shaped the final form of the output? Choices about prompt design, model selection, iteration, and curation can all potentially contribute to an affirmative answer. The difficulty is that the contribution must be more than mechanical or purely functional. A user who types a five-word prompt and accepts the first output has almost certainly not satisfied the threshold. A creative director who iterates through dozens of prompts, selects and edits outputs, and imposes a distinctive editorial voice on the result occupies far more defensible ground.
The distinction matters commercially. A business that cannot claim copyright in its AI-generated content has no basis to prevent competitors from reproducing it. The missed opportunity is real and, for content-intensive businesses, potentially severe.
The gap between statute and practice in Denmark
Danish copyright legislation has not been amended to address AI-generated works directly. The Kulturministeriet (Danish Ministry of Culture) has consulted on the issue, but no specific legislative reform has been enacted at the time of this analysis. Practitioners working in Denmark are therefore operating in a condition of genuine doctrinal uncertainty.
That uncertainty has a structural character. The statute requires human authorship. The practice of AI-assisted content creation introduces a spectrum of human involvement. At one end sits fully autonomous generation with minimal human input. At the other sits heavily curated, AI-assisted creation where the human contribution dominates. Danish courts will need to draw a line somewhere on that spectrum – and they have not yet done so clearly.
Several interpretive approaches are available. One holds that any meaningful human editorial judgment in the production chain is sufficient to ground copyright in the human contributor. A second, stricter approach insists that the human contribution must be evident in the final form of the work – not merely in the process that generated it. A third approach focuses on the contractual and organisational allocation of creative credit, asking who – within a defined team or enterprise – bears responsibility for the expressive choices made.
Each approach produces different results for the same AI-assisted workflow. Businesses that have not mapped their workflows against these interpretive options are carrying legal risk they may not have priced. For international clients considering Denmark as a content production or technology licensing hub, the absence of settled doctrine is itself a material due diligence consideration.
Related uncertainty extends to software liability for outputs that infringe third-party rights. If an AI system generates content that reproduces a protected work, the question of who bears liability. the developer. The deployer. Alternatively, the end user. is addressed only imperfectly by existing Danish tort and contract principles. Technology licensing agreements between AI developers and their customers increasingly attempt to allocate this risk contractually, but Danish courts have not yet validated any standard allocation model.
For a detailed review of intellectual property protection strategies in Denmark, including registration, enforcement, and licensing considerations, our dedicated service overview sets out the key procedural steps.
EU-level regulatory overlay: the AI Act and algorithmic accountability
The EU AI Act, which entered into force in 2024 and applies in Denmark as an EU Member State, introduces a layered set of obligations that intersect directly with the AI-generated works question. Its relevance to intellectual property is indirect but commercially significant.
The AI Act's transparency obligations require that certain categories of AI-generated content be disclosed as such. For businesses whose commercial proposition depends on the perceived authenticity or human authorship of their content, this creates a disclosure dilemma. Content that must legally be labelled as AI-generated may struggle to attract the same commercial premium as human-authored work. The tension between transparency law and commercial positioning is one that Danish businesses are only beginning to manage.
Beyond transparency, AI Act compliance requires deployers of general-purpose AI models to maintain technical documentation, conduct conformity assessments where applicable, and implement measures addressing algorithmic accountability. In the content production context, algorithmic accountability includes the ability to explain how a given output was generated – a requirement that intersects with copyright attribution questions. If a business cannot account for the generative process, it will struggle both to claim copyright protection and to defend against infringement allegations from third-party rights holders.
Danish supervisory authorities are in the process of designating the national market surveillance body responsible for AI Act enforcement. Until that designation is complete and enforcement practice develops, businesses must interpret the regulation's requirements in the absence of domestic guidance. The prudent approach is to apply the regulation's text directly, supplemented by guidance from the European AI Office where available.
The AI Act also addresses training data. Providers of general-purpose AI models must publish summaries of the copyrighted materials used to train their systems. This obligation has direct relevance to downstream users in Denmark. A business that licences a general-purpose AI system for content production should verify, through its technology licensing arrangements, that the model provider has satisfied its training data disclosure obligations. Failure to do so may expose the business to secondary liability claims if the model produces outputs that infringe third-party rights.
To explore how our firm advises on the full spectrum of AI regulation and content strategy, see our overview of AI and technology law services in Denmark.
Cross-border implications for European clients
Denmark does not operate its intellectual property system in isolation. As an EU Member State, Danish copyright law is shaped by harmonisation directives. The originality standard applied by Danish courts draws on the EU's autonomous concept of originality, developed through the case law of the EU-Domstolen (Court of Justice of the European Union). That concept – the author's own intellectual creation – is the same standard applied in Germany, France, the Netherlands, and other major European markets.
The practical consequence is that a copyright analysis conducted for Denmark will transfer, with adjustments, to most EU jurisdictions. Conversely, a business that structures its AI-assisted creative workflow to satisfy the human-authorship requirement in Denmark is likely building a defensible position across the EU. This portability of legal strategy is commercially valuable for businesses operating at scale across multiple European markets.
Where Denmark diverges from some EU neighbours is in the absence – so far – of court decisions that specifically engage with AI-generated outputs. German courts and French courts have addressed aspects of the question, developing nuanced positions on software-assisted creation that Danish practitioners monitor closely. The Højesteret (Supreme Court of Denmark) has not yet ruled on AI authorship directly. When it does, its reasoning is likely to engage with these European comparators.
For businesses with operations in both Denmark and Portugal. for example. Technology companies using Portugal's favourable tax conditions for IP holding structures. a comparative analysis of how each jurisdiction treats AI-generated works is a necessary part of IP portfolio planning. A parallel analysis of the Portuguese position is available in our deep analysis of AI-generated works and intellectual property in Portugal.
The cross-border dimension also arises in enforcement. A Danish rights holder whose AI-assisted content is reproduced without authorisation in another EU Member State must navigate the enforcement rules of the target jurisdiction. Where the competing copyright claim involves both a Danish originating work and a foreign reproducer, questions of applicable law and jurisdiction become material. Danish procedural rules on cross-border IP enforcement follow EU instruments, but the substantive analysis of whether the original work attracted protection remains anchored in Danish law.
For businesses holding digital services portfolios across Europe, the interaction between AI-generated content and platform liability rules adds a further layer. EU digital services legislation imposes notice-and-action obligations on platforms that host user-uploaded content. Where that content includes AI-generated material that may infringe third-party rights, the platform's response procedures must be calibrated to accommodate the ambiguity of the authorship question.
To receive an expert assessment of your AI content strategy in Denmark, contact us at info@ferrazwhitmore.com.
Strategic recommendations for international businesses
The doctrinal and regulatory uncertainty described above does not prevent effective legal planning. It does require that planning to be more deliberate than in settled areas of law. The following considerations define the terrain.
Map human creative contribution with precision. Businesses that rely on AI-generated content commercially should document the human creative decisions embedded in their workflows. Prompt design, iteration records, editorial selection criteria, and post-generation editing all potentially constitute human authorship contributions. A contemporaneous record of these decisions strengthens any future claim to copyright protection and is far easier to build prospectively than to reconstruct after a dispute arises.
Audit technology licensing arrangements. Every AI-assisted content business licences at least one foundational model. Those licence terms govern what the business can do with outputs, who bears liability for third-party infringement, and what training data the model provider has used. A licence that is silent on copyright ownership of outputs – or that allocates ownership to the model provider – may defeat the business's commercial proposition entirely. Danish contract law will interpret ambiguous licence terms against the background of general principles of construction, but the absence of specific AI-output provisions creates avoidable risk.
Build AI Act compliance into operations from the outset. Transparency disclosures, technical documentation, and conformity assessment processes are not bureaucratic inconveniences. They are the evidentiary infrastructure that demonstrates a business's compliance posture. In Denmark, where supervisory enforcement practice is still developing, early compliance investment positions a business favourably when the first enforcement actions are brought.
Consider the IP holding structure. For businesses generating significant volumes of AI-assisted content, the question of where copyright is held – and by which entity – has both commercial and tax dimensions. Denmark's IP holding environment, including its participation in EU IP box regimes, is relevant to the structuring decision. A rights structure built before the copyright ownership question is resolved may need to be unwound or reorganised when doctrine settles.
Monitor legislative developments actively. The Danish Ministry of Culture's ongoing consultation process may produce targeted amendments to copyright legislation. EU-level discussions on AI and copyright – including the ongoing review of harmonisation instruments – will shape the domestic Danish position. A business that monitors these developments can adapt its practices before competitors do. Capturing a first-mover advantage in a market where legal clarity, when it arrives, is likely to reward those who positioned correctly in advance.
For a tailored strategy on AI-generated content protection in Denmark, reach out to info@ferrazwhitmore.com.
Outlook: where Danish law is heading
The trajectory of Danish intellectual property doctrine on AI-generated works is not difficult to discern, even if the precise destination remains uncertain. Several vectors are converging.
First, the EU legislative agenda will drive change. The European Commission's work on copyright in the context of AI – including targeted consultations and potential amendments to harmonisation instruments – will constrain Denmark's domestic options. Any Danish legislative reform is likely to operate within parameters set at EU level, meaning that the relevant lobbying and standard-setting activity happens primarily in Brussels, not Copenhagen.
Second, court decisions will accumulate. The Maritime and Commercial Court and, ultimately, the Supreme Court of Denmark will be called upon to rule on AI authorship questions as commercial disputes reach them. The first significant ruling – whenever it comes – will set a reference point that reshapes market practice rapidly. Businesses that have not prepared for either outcome – copyright protected or not – are exposed to a binary risk.
Third, the AI Act's implementation timeline will force operational changes regardless of how copyright doctrine develops. The transparency, documentation, and accountability obligations imposed by the regulation apply to AI-generated content broadly, not only to content with copyright ambiguity. Compliance with the AI Act will, as a by-product, generate much of the documentation that businesses need to support copyright claims.
Fourth, industry practice in Denmark is already moving toward hybrid human-AI workflows that are designed to satisfy the human-authorship threshold. Creative directors are being assigned formal responsibility for AI output selection. Editorial protocols are being documented. These practices, which are emerging from commercial necessity rather than legal compulsion, are likely to become the de facto standard that courts recognise as sufficient for copyright protection.
The businesses best positioned to capture value from AI-generated content in Denmark are those that treat legal structuring as a competitive advantage rather than a compliance burden. The window for establishing strong positions – before doctrine settles and competitors follow – is open now.
Frequently asked questions
Q: Can an AI system hold copyright in Denmark?
A: No. Danish intellectual property legislation conditions copyright protection on a human author who exercises creative choices. An AI system has no legal personality and cannot hold rights. The relevant question is whether a human contributor – the developer, the operator, or the user – made sufficiently original creative decisions to attract protection in their own name.
Q: How does the EU AI Act affect AI-generated content businesses operating in Denmark?
A: The EU AI Act introduces transparency and documentation obligations that apply directly in Denmark. Businesses deploying general-purpose AI systems must maintain technical documentation, disclose AI-generated content where required, and observe algorithmic accountability standards. Non-compliance carries significant administrative penalties under the regulation's enforcement regime, which Danish supervisory authorities will apply from the relevant phase-in dates.
Q: What practical steps should a technology company take before commercialising AI-generated content in Denmark?
A: A technology company should first audit the human contribution embedded in its AI workflow to assess whether copyright protection is available. It should then review its technology licensing arrangements – particularly training data licences – to confirm they permit commercial output. Finally, it should map its obligations under AI Act compliance requirements, including any transparency disclosures and documentation duties that apply to its specific use case. Engaging a lawyer in Denmark with cross-border AI experience is advisable before launch.
About Ferraz & Whitmore
Ferraz & Whitmore is an international law firm based in Lisbon, advising business clients across 46 jurisdictions. Our IP and technology law practice advises technology companies, content platforms, and institutional investors on AI-generated works, intellectual property protection, and AI Act compliance across European and international markets. As a law firm in Denmark and across the EU, we combine civil law precision with common law enforcement strategy to support clients managing AI content portfolios at scale. Our attorneys have advised on technology licensing, digital services regulation, and cross-border IP enforcement in both civil law and common law systems. The firm's Lisbon base provides direct access to Portuguese and EU regulatory regimes, while our common law expertise supports enforcement strategies in English-speaking jurisdictions. We are a member of leading international legal associations and participate in cross-border practice groups focused on AI regulation and intellectual property. To discuss your AI content strategy in Denmark or across Europe, contact us at info@ferrazwhitmore.com.
Disclaimer: This publication is provided for informational purposes only and does not constitute legal advice. The information herein should not be relied upon as a substitute for professional legal counsel tailored to your specific circumstances. Ferraz & Whitmore assumes no liability for actions taken or not taken based on the contents of this material. For advice regarding your particular situation, please contact info@ferrazwhitmore.com.