Who owns the copyright in generative AI output?
The use of generative artificial intelligence (“AI”) platforms has grown at an unprecedented rate over the past five years with Microsoft estimating that one in six people worldwide use AI tools. Applications such as ChatGPT are increasingly used to produce songs, images, books, and cartoons.
This article considers two key legal questions under English and Welsh law:
- Can copyright subsist in works created by generative AI?
- If so, who owns that copyright?
Copyright law
Copyright is a form of intellectual property that gives the owner the exclusive right to control how a work is used or copied.
In England and Wales, copyright protection is primarily governed by the Copyright, Designs and Patents Act 1988 (“1988 Act”). In certain categories of work copyright arises automatically without any need for registration, for example, literary, dramatic, musical, and artistic works. However, the work must be original.
To be considered ‘original’, the work must be the author’s own intellectual creation, rather than a copy of an existing work. However, originality is not required for all categories of copyright works.
In the case of films, sound recordings, and broadcasts, copyright arises upon creation, irrespective of whether the content is original. In those cases, copyright protection reflects the skill, labour, investment, and organisation involved in producing the work.
As a general rule, the author or creator of a work is the first owner of the copyright, subject to limited exceptions.
Copyright and AI‑generated works
The courts of England and Wales are yet to rule directly on the copyright status of works produced by generative AI systems.
Overseas courts have, however, adopted a restrictive approach. In the United States case of Thaler v Perlmutter (2023), a federal district court upheld the US Copyright Office’s refusal to register a wholly AI‑generated work. The Court confirmed that copyright protection required human authorship.
Similarly, South Korean courts have declined to recognise copyright in AI‑generated works unless the output clearly conveys human thoughts and emotions.
Computer‑generated works under the laws of England and Wales
Section 9(3) of the 1988 Act contains a specific rule for computer‑generated works, providing that:
“In the case of a literary, dramatic, musical or artistic work which is computer‑generated, the author shall be taken to be the person by whom the arrangements necessary for the creation of the work are undertaken.”
A “computer‑generated work” is defined in section 178 as a work that is:
“generated by computer in circumstances such that there is no human author of the work.”
Although drafted in the late 1980s, this provision may capture many AI‑generated outputs, particularly where no human intellectual creation is expressed in the work itself.
In its March 2026 Report on Copyright and Artificial Intelligence, , the Government questioned whether section 9(3) remains appropriate in the age of modern AI.
However, no decision has yet been taken to amend or repeal this provision.
AI‑generated works and “non‑original” works?
Films and sound recordings are defined statutory categories that do not require originality. AI‑generated content, by contrast, often adopts the form of literary, musical, or artistic works. This may dispense with the requirement of human authorship through section 9(3).
This distinction between different types of works is also reflected in the duration of protection. Under section 12(7) of the 1988 Act, copyright in computer‑generated works lasts 50 years from the end of the year of creation. In the case of non‑computer‑generated works such as literary, dramatic, musical, and artistic works, copyright usually lasts for the author’s life plus 70 years.
Who owns copyright in AI‑generated works?
According to section 9(3), in the case of literary, dramatic, musical or artistic works (which are computer generated), “the author shall be taken to be the person by whom the arrangements necessary for the creation of the work are undertaken.”
However, this has not yet been tested by the courts of England and Wales in the context of modern generative AI. Depending on the facts, potential candidates for authorship may include:
- The user who prompts or directs the AI platform;
- The developer or operator of the AI platform; or
- An employer, where the relevant arrangements are carried out by an employee in the course of their employment (in which case section 11(2) of the 1988 Act vests ownership in the employer, subject to any agreement to the contrary).
There has been some developing caselaw abroad.
In the Chinese cases of Tencent v Yingxun (2019) and Li Yunkai v Liu Yuanchun (2023), the courts have favoured the user by recognising copyright in AI‑assisted content where there was sufficient human involvement.
Earlier UK cases involving computer‑generated works, such as Nova Productions Ltd v Mazooma Games Ltd (2007), did not determine authorship under section 9(3) and therefore provide limited guidance.
It is arguable that, had Parliament intended the creator of the AI software to be treated as the author, it would have said so expressly. Section 9(3) appears to avoid any reference to programmers or software developers.
Practical tips
If you create content using generative AI, it is essential to review the platform’s terms and conditions. In practice, contractual terms may determine ownership.
For example, ChatGPT’s terms currently state:
“As between you and OpenAI, and to the extent permitted by applicable law, you (a) retain your ownership rights in Input and (b) own the Output. We hereby assign to you all our right, title, and interest, if any, in and to Output.”
Such provisions can be commercially significant, but terms vary between providers and are often open to interpretation or dispute.
Conclusion
Until the courts of England and Wales or Parliament itself provide further clarity, the legal position regarding generative AI remains unsettled.
At this stage, there is no definitive answer as to who owns copyright in AI‑generated works as between the AI software provider and the user.
The contents of this article are intended for general information purposes only and shall not be deemed to be, or constitute legal advice. We cannot accept responsibility for any loss as a result of acts or omissions taken in respect of this article.