As AI technology develops, we are now firmly in the age of non-humans authoring literary content which might be worthy of protection under intellectual property laws.
In this blog post, we consider the sophisticated artificial intelligence program ChatGPT, and the copyright issues that may arise in context of the evolving copyright legislation in the United Kingdom.
In the UK, copyright protection is largely governed by the Copyright, Designs and Patents Act 1988 (CDPA). Copyright protection extends to types of creative endeavours that are recorded in some way, which includes original literary, dramatic, musical or artistic works. The general rule is that the author/creator of the work is the first owner, however this legislation was created with human authors in mind as the creators. In the advent of new technologies such as AI, the position is not as clear cut in relation to i) ownership of the output; and ii) whether it qualifies for copyright protection at all. We consider these issues in greater depth below.
ChatGPT was developed by Microsoft-backed company OpenAI and follows in the wake of Dall-E, OpenAI’s AI generator of digital images from text prompts. ChatGPT is a chatbot language model that has been trained on large amounts of data that includes text, books and articles, amongst other sources. It uses this ‘deep learning’ to respond to questions based on an AI-enhanced algorithm. ChatGPT has been trained to respond in a conversational and human-like manner and has been available for free public testing since 30 November 2022. Reports from ChatGPT users have been largely positive, who have asked the chatbot to produce essays, summarise history, dispense basic medical advice, write poems, menus and jokes for example.
This is potentially a complex area at the present time.
Under UK law, in order for a work to qualify for copyright protection, the work needs to be original. This originality hurdle is often considered to be something of a low bar. However, in the context of AI-generated context it could be said in some contexts that there is no (human) originality at all and thus such AI-works are undeserving of protection.
A further complication is that the concept of originality is subject to uncertainty and potential change. This is because the current ‘originality’ requirement (derived from EU case law) is that that copyright will subsist in a work if it is the “author’s own intellectual creation” that is an expression of that creation. This contrasts with the historical position under UK law, which has deemed a ‘work’ (defined under the CDPA) to be to be sufficiently original, if it is the result and application of an author’s “labour, skill or effort”, which is arguably a lower threshold. Furthermore, post-Brexit, this is perhaps an area which is ripe for potential divergence now that English courts are no longer bound by EU case law.
AI tools are also utilised in different ways, which may render a different outcome in terms of copyright protection. For example, in situations where the AI is a merely facilitative tool, then perhaps the human author would be considered to be deserving of owning a copyright. This is because, where the AI is simply a tool, there is scope for the human author to exercise his/her own creative ability to make free and creative choices in the ultimate output of the literary work for example, to meet the ‘originality’ requirement.
Conversely, where content generated by AI-systems like ChatGPT become more autonomous and require less and less human intervention, perhaps the human input falls away together with the entitlement to copyright.
A further issue is whether the output generated by the AI is truly ‘original’ based on the fact that the output is generated by the AI’s deep learning derived from pre-existing data.
The UK is alive to these issues in respect of intellectual property rights and AI, with an aim to ensure that legislation evolves in tandem with technological advances. In September 2020, the government launched an initial call for views which ran from 7 September – 30 November 2020, specifically in relation to legislative policy around patents, data mining and copyright. In terms of copyright, the proposals announced in March 2021 covered i) review of licensing around copyright works for use with AI; and ii) further consultation on whether ownership in original works should be limited to human creations (including AI-assisted creations). A further consultation ran from 20 October 2021 – 7 January 2022 (responses can be viewed here) where it was concluded, inter alia, that the use of AI to generate creative content is still in its early stages and consequently, there would be no imminent legislative changes in relation to computer generated works under the CDPA. However, the IPO will continue to monitor the impact of existing protection as AI develops and welcomes any new evidence from stakeholders.
Waterfront asked ChatGPT whether its output could be eligible for copyright protection. The response was as follows:
“In general, copyright protection applies to original creative works that are fixed in a tangible form, such as a written document or a computer file. This means that if ChatGPT generates a creative work that is saved in some way, such as by being written down or saved to a file, it may be eligible for copyright protection.
However, the extent to which a work generated by ChatGPT (or any other AI system) is eligible for copyright protection can depend on various factors, including the level of originality and creativity involved in the creation of the work and the level of human input or intervention in the process.
In the UK, the law regarding copyright protection for works created by AI is still developing, and there is currently no specific provision in the Copyright, Designs and Patents Act 1988 or related legislation that addresses this issue. In general, however, it is likely that a work created by AI would need to meet the same standards of originality and creativity as a work created by a human in order to be eligible for copyright protection.
It is important to note that this is a complex and evolving area of law, and the specific copyright implications of a work generated by ChatGPT or any other AI system may depend on the specific circumstances of the case. If you have specific questions or concerns about the copyright status of a work created by ChatGPT or any other AI system, it is advisable to seek legal advice”
As summarised quite accurately by ChatGPT, the position around ownership of AI generated content remains uncertain.
If your business is generating or using AI content, and until further clarity is provided by legislative development, it may be appropriate to consider whether contractual provisions might be required to protect your position. Furthermore, and to quote ChatGPT, if you have any questions or concerns, “it is advisable to seek legal advice”.
 S1.1(1)(a) CDPA
 Infopaq C5/-08 (2009); Brompton Bicycle C-833/18 (2020)
A recent EU trade mark application for the word mark, PUT PUTIN IN, has been refused by the European Union Intellectual Property Office on the grounds of being contrary to public policy or to accepted principles of morality. While a fairly straightforward decision, this is a timely reminder…
Late yesterday UK time, it was reported that a lawyer for Twitter had sent a letter to Meta CEO Mark Zuckerberg complaining about Meta’s new Threads app. Twitter claimed that it “has serious concerns that Meta Platforms (Meta) has engaged in systematic, wilful and unlawful misappropriation of Twitter’s trade secrets and other intellectual property”.