Copyright and artificial intelligence. What a new EU briefing says about authorship in the age of generative AI

As generative artificial intelligence becomes a common tool for creating text, images, music and video, lawmakers and courts around the world are being forced to reconsider long-standing ideas about authorship and creativity. A December 2025 briefing by the European Parliamentary Research Service examines how different legal systems are responding to this challenge, with a particular focus on the European Union and selected countries outside it EPRS_BRI(2025)782585_EN[1].

What is the problem copyright law is facing?

Copyright exists to protect human creativity. It gives authors control over how their work is used and allows them to earn a living from it. Generative AI complicates this model because it can produce content that looks creative, even when a human’s role is limited to typing prompts or selecting outputs.

This raises practical questions. If an AI system explains a poem, creates an image or composes music, who, if anyone, is the author? Is it the person who wrote the prompt, the developer of the AI, or no one at all?

The European Union’s approach. Humans at the centre

The briefing explains that the European Union does not have specific rules on the copyrightability of AI-generated works. However, existing EU law and court rulings point strongly toward a human-centred approach.

The Court of Justice of the European Union has repeatedly held that a work must reflect the author’s “own intellectual creation”. In simple terms, copyright applies only when a human makes free and creative choices that shape the final result. This makes it unlikely that content generated entirely by AI, with little or no human input, would qualify for copyright protection in the EU.

What remains unclear is how much human involvement is enough when AI tools are used as part of the creative process. Editing, selecting outputs, or arranging AI-generated material may count, but there is no clear threshold yet.

How EU countries and the European Parliament see it

According to the briefing, EU member states broadly agree that “significant” human input is required for copyright protection. Italy clarified this in 2025 by explicitly stating that works created with AI tools can be protected, as long as they are the result of human intellectual effort.

The European Parliament has taken a similar position. It has consistently argued that purely AI-generated content should not be copyrightable, while calling for clearer criteria to assess human authorship in AI-assisted creations.

Different answers outside the EU

The briefing shows that there is no global consensus.

In the United States, copyright law remains firmly human-centric. Courts and the US Copyright Office have ruled that prompts alone are not enough to make someone an author, although human-created elements, such as text or creative arrangement, can still be protected.

Some common-law countries, including the United Kingdom and Ireland, take a different approach on paper. Their laws allow copyright for “computer-generated works” and assign authorship to the person who made the necessary arrangements. However, these rules were written decades ago and have not yet been fully tested in modern AI cases.

China’s courts have issued mixed rulings. In some cases, they have recognised copyright where users showed substantial creative effort in selecting, refining and editing AI outputs. In others, they have rejected claims due to insufficient originality.

Ukraine stands out for introducing a special, or ‘sui generis’, right for AI-generated images. This offers limited protection without treating AI outputs as traditional copyrighted works. Most EU countries, however, have rejected this approach, warning it could weaken the position of human creators.

Why this matters

As the briefing concludes, differing national approaches create legal uncertainty. The same AI-assisted work may be protected in one country and rejected in another. This complicates licensing, enforcement and cross-border use of creative content.

For now, most legal systems still insist on meaningful human involvement. But as AI tools become more powerful and harder to distinguish from human creativity, courts and policymakers will face increasing pressure to clarify where authorship begins and ends. According to the European Parliamentary Research Service, finding a balanced and coordinated approach may be essential to protect both innovation and the livelihoods of human creators in the years ahead