Navigating Copyright and Copyright Infringement In The World of AI

UK copyright law is primarily governed by the Copyright, Designs and Patents Act (CDPA) 1988. Section 9(3) of the CDPA provides scope for the inclusion of AI-generated art within the CDPA as this provision stipulates that:

"In the case of...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".

The nature of the 'prompting system' for AI whereby the user 'prompts' the AI system to produce a result consistent with the user's request provides greater legal difficulty concerning s9(3) CDPA as it may be unclear exactly who undertook the necessary arrangements for the creation of the work.

Since joint authorship between a human user and AI system is impossible under the CDPA, the degree and extent of the 'prompting' will likely determine the work's author and subsequent copyright owner.

However, it remains unclear whether user intellectual effort via extensive prompting of the AI system or the use of AI as a source of inspiration for further user intellectual effort will be sufficient to satisfy the "personal touch" requirement for copyright protection.

Due to the current uncertain legal position, documentation of all prompting and subsequent human alterations, for example, through image editing software, is imperative as cases are likely to be decided on a case-by-case basis.

The UK Government's consultation in December 2024, part of the broader AI Opportunities Action Plan, seeks to clarify such ambiguities in copyright law.

This initiative reflects the Government's goal to ensure that the legal landscape supports innovation while providing clear guidance for users and developers navigating AI-generated works.

Infringement by training the AI system

The first infringement issue AI presents relates to the significant volume of data which must be 'scraped' from the internet to train the AI system.

Copyright infringement requires the use of all or a substantial part of copyright works without authorisation from the owner.

In the context of AI, the recent Getty Images (US) v Stability AI Ltd [2023] EWHC 3090 (Ch) case involved an allegation that the data used to train Stability AI's system included Getty Images' work as Stability AI sought a reverse summary judgment and strike out of Getty's claim.

Although the case will be heard at trial later this year, the Getty Images case demonstrates the potential liability for copyright infringement for AI developers since users, with sufficiently precise prompts, could produce an image which included the Claimant's logo.

The Government has underscored the importance of ensuring AI systems adhere to ethical principles, including transparency in training data. The AI Opportunities Action Plan highlights that a trustworthy AI ecosystem depends on documenting and ethically sourcing data used in model development.

Infringement by the user

Moreover, the current position for copyright infringement by individuals and companies prompting the AI system ('users') is equally unclear.

Given that copyright infringement by users remains untested in Court, significant caution should be exercised by users who are aware of a particular artist's work and enter prompts into the AI system to create similar work in the style of a particular artist.

Similarly to the issue of authorship, and given the legal uncertainty at this stage, the documentation of all prompts entered by a user could prove vital in copyright infringement proceedings.

This aligns with the UK Government's ongoing plans to support responsible AI use.

As noted in the BBC, fostering public trust in AI will require that ethical practices are promoted and enforced through transparent regulations and practical guidance for individuals and organisations.