In a previous article, ‘Your Intellectual Property and AI‘, we discussed Generative AI’s potential infringement of existing intellectual property rights. This article focuses on copyright protection relating to works created by Generative AI. We consider arguments relating to whether copyright law in the UK should be reformed to provide new rights to works created by Generative AI.
Copyright and Generative AI – an introduction
Copyright is an intangible property right that allows a piece of work to be reproduced in any form.
The Copyright, Designs and Patents Act 1988 (CDPA 1988) governs copyright in the UK and provides that copyright will apply to the following types of works:
- literary, dramatic, musical or artistic works
- sound recordings, films or broadcasts
- the typographical arrangement of published editions (i.e. the style, composition, layout and general appearance of a page of a published work)
Copyright is not registerable in the UK. It arises automatically, meaning no paperwork needs to be filed or fees paid for copyright to be established in a work.
CDPA 1988 provides that copyright arises in an original work. It is generally accepted that the UK follows the EU’s test of what is considered “original”, namely that a work is original when it is “the author’s own intellectual creation”.
Works produced by Generative AI (which fall into one of the categories listed above) are, to an extent, protected under CDPA 1988. It is not clear yet how this protection can be assumed safely to apply. However, UK statute does not currently expressly cover this, and there are likely to be issues establishing who the author of AI work is for the purpose of effective copyright protection if this is not immediately clear.
Generative AI and copyright authorship – the issue
Section 9(3) of CDPA 1988 is key in relation to the Generative AI authorship issue. The section provides 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’.
To establish copyright in a computer-generated work, the author is taken to be the person who undertakes and makes the necessary arrangements for the creation of the work to take place.
Generative AI which can create new content independently is new. For example, AI chatbots can produce new work in response to very few text prompts given by a user.
This presents challenges in determining who has undertaken the necessary arrangements to create the work produced by the Generative AI. Was it the developer of the AI chatbot? Was it the user of the AI chatbot? Where this is unclear, it is difficult to establish who the author of the work is for the purpose of establishing copyright ownership and so protection. Even if it is possible to identify an author, it is still necessary to establish whether the work can be considered the author’s own original intellectual creation which is required to establish effective copyright protection.
Copyright law has not yet been tested with respect to Generative AI created work. In the media and entertainment sector, it is a widely held view that work created by Generative AI should not be copyright protected. An update of UK copyright law clarifying the position on works created by Generative AI is expected.
Things to keep on your radar
Reform of copyright law – resolving the authorship issue
In Nova Productions Ltd v Mazooma Games Ltd [1](Nova), the judge in this case held that the author of works in a video game was the game developer,[2] and the player was not the author as he had not ‘undertaken any of the arrangements necessary for the creation of the frame images. All he has done is to play the game’.[3]
Similarities and Differences Between Video Games and Generative AI
There are similarities between video games and Generative AI. For example, users of both technologies must input prompts to instigate a response and the developers of both technologies invest significant time and skill in creating the experience of using these technologies. In the future. the courts could apply the decision in Nova to determine the position on copyright protection of Generative AI-produced works, by confirming developers of Generative AI are the authors of a work for the purpose of copyright protection, not users.
It is relevant though that there are significant process capability and development differences between the production of video games and Generative AI. The development of video games involves programming, using predefined logic systems and rules to produce outcomes within each game. Video game developers undertake the necessary arrangements to create visual, audio, and literary work in producing a video game. Generative AI developers do not use task specific programming and instead use large data sets to “train” and so facilitate Generative AI functionality.
Because of the different processes in creating a video game and Generative AI functionality and resulting works, it is not clear at this stage that the principles established in Nova can readily be applied in the context of a Generative AI dispute. It will be helpful to see how the UK courts address and determine these challenges.
Potential Amendments and Legal Interpretations
Parliament could amend CDPA 1988, or the courts could expand the definition of ‘person’ so that the technology itself is capable of being held to be the author of a work for the purpose of copyright protection, though this seems unlikely.
In the UK, an artist is the author of a work for the purposes of copyright, where, for example, the artist is commissioned to produce a piece of art for a client (unless there has been an agreement between the client and the artist reversing this position).
In March 2023, the UK Supreme Court established that AI could not be an inventor for the purposes of a patent.[4] This principle, suggests then that, like an inventor, a ‘person’ under CDPA 1988 is unlikely to be determined as including a technology as an author of a work. This would be consistent with current developments in the US and EU.
Conclusion
As with the creation of any new technology, the law takes a while to catch up with the new situation which means we will be in a period of uncertainty in the short term. While the position continues to develop and work through, we must rely on established current law and precedent applying it to the new ownership and authorship uncertainty. The challenge will be to gather adequate evidence to be able to confidently defend copyright infringement and enforce copyright protection rights in the context of the current uncertainty in relation to the authorship and ownership of works created by Generative AI. New law and regulation is likely to be required to assist in establishing a platform for the protection of copyright in this area.
Our Corporate & Commercial team can discuss these AI risks with you further. They can assist your company in preparing and implementing AI policies to mitigate these risks.
Contact us today by emailing online.enquiries@la-law.com or calling 01202 786188.