Who Owns AI-Generated Works? The UK’s Copyright Dilemma from 1988 to 2026
Who is the author of an image, a piece of music, or a written text produced by generative artificial intelligence? Is it the developer of the AI system, the person who entered the prompt, or someone else responsible for making the production possible?
These questions have become increasingly relevant with the widespread use of generative AI. Yet the legal problem is not entirely new.
The United Kingdom introduced a specific copyright provision for computer-generated works in 1988, long before today’s generative AI systems existed. Rather than recognising computers as legal authors, British legislation adopted an unusual solution: it attributed authorship to the person responsible for making the arrangements necessary for the creation of the work.
Almost four decades later, this provision has become a source of significant legal uncertainty. In March 2026, the UK government published a report proposing the removal of the specific copyright protection for works generated without a human author.
The history of this provision raises a fundamental question: Can copyright law solve the problem of AI-generated works simply by assigning authorship to a person?
A copyright solution developed before generative AI
The Copyright, Designs and Patents Act 1988 (CDPA) contains three important provisions concerning computer-generated works.
Section 178 defines a computer-generated work as one generated by a computer in circumstances where there is no human author.
Section 9(3) provides that, for a computer-generated literary, dramatic, musical or artistic work, the author is deemed to be the person who undertook the arrangements necessary for its creation.
Section 12(7) establishes a specific term of protection: 50 years from the end of the calendar year in which the work was made.
The crucial distinction is that using a computer does not necessarily make a work computer-generated within the meaning of the CDPA.
An artist creating a digital illustration, a composer using music-production software, or a photographer editing an image may still be the human author of the resulting work. The special provision becomes relevant where the work has no human author.
What makes section 9(3) particularly interesting is that authorship is not necessarily attributed to the person who made the creative choices embodied in the work.
Instead, the law identifies a person responsible for arranging the circumstances of its production.
The provision does not automatically identify the programmer, the owner of the system or its end user as the author. Determining who made the necessary arrangements may depend on the facts of each case.
Historically, the provision was intended to address uncertainty about the legal status of outputs produced through increasingly automated computer processes. It also reflected concerns about encouraging investment in computing technology.
Rather than establishing a separate sui generis right, the UK chose to accommodate these outputs within its existing copyright framework.
The historical British approach to originality
To understand why such a provision was introduced in 1988, it is useful to consider the development of originality in British copyright law.
Traditionally, British copyright jurisprudence attached considerable importance to the skill, labour and judgment involved in producing a work.
This approach, commonly associated with the expression sweat of the brow, was less dependent on identifying an author’s personal creative choices than the originality standard subsequently developed in European copyright law.
The often-cited expression what is worth copying is worth protecting illustrates an important aspect of that historical tradition, although it should not be understood as eliminating the requirement of originality.
Against this background, the idea of protecting computer-generated outputs through a statutory attribution of authorship was not entirely inconsistent with the historical logic of British copyright.
However, the development of EU copyright law substantially changed the legal landscape.
Under the originality standard established by the Court of Justice of the European Union, copyright protection requires a work to be the author’s own intellectual creation.
This test is associated with the author’s free and creative choices and the expression of human intellectual creativity.
The influence of this standard on British copyright law has continued beyond Brexit.
In THJ Systems Ltd v Sheridan [2023] EWCA Civ 1354, the Court of Appeal of England and Wales confirmed that the relevant originality assessment was based on the author’s own intellectual creation rather than the traditional skill and labour approach.
This development creates a difficulty for computer-generated works.
Can a work without a human author be original?
Section 9(3) of the CDPA provides a legal mechanism for attributing authorship to a person where a work has no human author.
However, identifying an author and establishing that a work qualifies for copyright protection are separate legal questions.
The fact that legislation deems someone to be the author does not automatically establish that the output satisfies the originality requirement.
This brings us to the central contradiction.
If originality requires an author’s own intellectual creation, reflecting free and creative human choices, how can a work generated without any human author satisfy that requirement?
On one hand, section 9(3) expressly contemplates works without human authors. On the other, the contemporary originality standard is closely associated with human creativity.
One possible interpretation would be to apply a different understanding of originality to computer-generated works, giving effect to Parliament’s original intention.
Yet it cannot be said that such an interpretation has been definitively established under current UK law.
The UK government’s March 2026 report explicitly acknowledges this legal uncertainty.
The report suggests that a court would be unlikely to interpret the legislation in a way that rendered the provision entirely ineffective, given Parliament’s evident intention. Nevertheless, there remains uncertainty, in the absence of definitive case law, about how an original work without a human creator should be identified.
Brexit has not automatically resolved this problem. Although UK courts may have greater scope to develop copyright doctrine independently, the originality standard derived from EU law continues to influence domestic jurisprudence.
Applying the 1988 provision to modern generative AI
The technology envisaged by the legislators in 1988 was considerably different from modern generative AI.
Today, the degree and nature of human involvement in AI-assisted creation can vary substantially.
Consider a graphic designer who generates images using an AI system, selects particular elements, modifies them, and arranges them into a new composition through independent creative decisions.
In such circumstances, the resulting work may qualify for ordinary copyright protection on the basis of the designer’s own creative contribution.
Now consider a user who enters a simple textual prompt into a generative AI system and receives an image without exercising meaningful creative control over its expressive features.
The legal analysis may be substantially different.
The UK government’s 2024 consultation and March 2026 report suggest that, in the case of general-purpose generative AI responding to a prompt, the person entering the prompt would usually be regarded as the person making the necessary arrangements under section 9(3).
However, this policy interpretation should not be confused with a definitive judicial rule applicable to every AI-generated output.
Nor does it mean that every person who enters a prompt automatically obtains copyright in the resulting material.
The question of who made the necessary arrangements remains distinct from whether the resulting output meets the requirements for copyright protection.
The distinction between AI-assisted and AI-generated works is therefore central to the debate.
March 2026: The UK government proposes removing the protection
In 2022, following an earlier consultation, the UK government decided to retain the existing protection for computer-generated works while keeping it under review.
The rapid development of generative AI subsequently prompted a reconsideration of that position.
A new consultation launched in December 2024 examined whether the existing provisions should be retained, clarified, reformed or removed.
On 18 March 2026, the government published its Report on Copyright and Artificial Intelligence, expressing a preference for removing the specific copyright protection available to computer-generated works without human authors.
Several considerations informed this position.
First, the available evidence does not demonstrate that the existing protection has materially encouraged investment, innovation or the production of computer-generated works.
Second, uncertainty remains about the relationship between statutory attribution of authorship and the requirement of originality.
Third, extensive copyright protection for automatically generated outputs may affect markets for human-created works, particularly where large quantities of AI-generated content compete with human creative production.
The consultation responses also revealed substantial support for removing the existing special protection, especially among respondents from the creative industries.
Nevertheless, it is important to distinguish a policy proposal from an enacted legislative amendment.
As of October 2026, the government’s proposal should not be treated as the repeal of section 9(3) itself. The report expresses a preferred policy direction, not a completed legislative change.
The UK has also considered removing comparable provisions for computer-generated designs without human authors, reflecting similar concerns within design law.
What does the British experience mean for copyright reform?
The British approach provides a useful perspective for jurisdictions considering whether and how AI-generated outputs should receive legal protection.
Under Turkish copyright law, the concept of a protected work is based on the requirement that it reflect the individual character of its author (sahibinin hususiyeti). The Law on Intellectual and Artistic Works, Law No. 5846, does not contain a general provision equivalent to section 9(3) of the CDPA for attributing authorship of wholly computer-generated works.
Accordingly, any proposal to protect outputs generated without human creativity should first identify the underlying legal interest that requires protection.
Is the purpose to reward creative intellectual activity? Or is it to encourage technological investment and protect the economic value of automated production?
These objectives are not necessarily identical, and they do not inevitably require the same legal mechanism.
If protection of wholly AI-generated outputs is considered economically justified, I do not believe that the only available option is to accommodate those outputs within the traditional concept of copyright authorship.
A separate, limited sui generis regime could also be considered.
However, such a regime would require careful consideration of its purpose, scope, duration, economic justification and effects on the public’s ability to access and reuse generated material.
The existence of economic value alone does not necessarily justify copyright protection.
In my view, the British experience illustrates that statutory attribution of authorship cannot, by itself, resolve the fundamental questions surrounding AI-generated works.
Perhaps the more important question is not who should be recognised as the author of an AI-generated output, but rather why, to what extent, and under which legal framework an output created without human intellectual creativity should be protected at all.
The distance between the UK’s 1988 legislative solution and its proposed reform in 2026 demonstrates why that distinction matters.
References
- Copyright, Designs and Patents Act 1988, particularly sections 9(3), 12(7) and 178.
- UK Government, Copyright and Artificial Intelligence Consultation (2024).
- UK Government, Report on Copyright and Artificial Intelligence (18 March 2026), particularly Section I.
- THJ Systems Ltd v Sheridan [2023] EWCA Civ 1354.
- UK Intellectual Property Office, Consultation on Changes to the UK Designs Framework.