Reading time: five minutes

Trained on vast datasets, Generative Artificial Intelligence models can produce impressively sophisticated outputs − some of which are perhaps indistinguishable from human-created works. However, these new tools raise significant legal quandaries, particularly for copyright law.

Interested in finding out more about AI? I’ve already written on GenAI concerns, as well as more specifically about ghostbots and deepfakes. 

When I first became aware of GenAI a few years ago, I was captivated by the possibilities, and the potential for development. However, as a then first year law student, I was also beginning to consider the myriad questions it raised − could a program really create art or written works of any merit at all, let alone those comparable to human creations? GenAI has come leaps and bounds since then, and this curiosity ultimately led me to write my dissertation this year on one particularly thorny legal issue − can GenAI be considered sufficiently original for the purposes of copyright authorship? It must be acknowledged that this topic can’t be fully explored in a 10,000 word dissertation, but I shall do my best in this article to briefly explain the core issues and principles of GenAI and copyright authorship. 

Defining originality

For UK copyright law, originality of works is a fundamental requirement for protection, requiring a work to originate from the author and not be a mere copy or derivation of another. In the UK, this has evolved from a standard of "skill, labour and judgement" to the "own intellectual creation" standard set by the EU Infopaq case. The EU standard reinforces the need for human involvement and creativity. This prompts questions about how to handle works created by AI with minimal human input. 

GenAI works challenge our typical notions of originality because they are ‘new’ works created from training data, which are existing works. The key issue with GenAI outputs is whether these works can be considered original. The Infopaq standard requires that the work reflects the author's personality and is the result of the author's free and creative choices. Because the AI’s choices are algorithmically determined, and the outputs are based upon its training data, scholars dispute whether these works are original or simply derivative. 

It should be acknowledged that humans don’t create in a vacuum − arguably everything we do, say, and create is based on our experiences, from things we’ve seen, read or heard. However, while human creators can bring their personal touch to their works, the copyrightability of GenAI works (and the authorship of these works) depends on the AI's ability to generate something truly original from its training data.

AI authorship

Authorship in copyright law traditionally implies human creativity. Section 9(3) of the UK's Copyright, Designs, and Patents Act 1988 (CDPA) recognizes computer-generated works, assigning authorship to the person “by whom the arrangements necessary for the creation of the work are undertaken”. This provision was relatively forward-thinking and some academics have praised its simplicity – but it’s being tested by the advancing capabilities of modern GenAI and may not be fully equipped to address issues of AI authorship.

The CDPA suggests that the human who sets up and prompts the AI could be considered the author. However, this definition raises questions about the extent of human involvement required to qualify as an author.

There’s a tumultuous debate on whether AI can (or indeed should) be granted authorship. Some scholars argue for the AI's legal personality and authorship. One suggestion is that if agents, and even “incorporated legal entities”, can be authors under the CDPA, logically so too could sophisticated AI systems.

Many scholars argue against AI authorship. Cases like the Naruto monkey selfie reinforce the necessity of human involvement in creative processes for copyright eligibility. In this case, a monkey (dubbed Naruto for the dispute) took photos using a photographer's camera. The court ruled that non-human animals cannot hold copyrights or be authors, but the photographer was also refused rights over the images. This case underscores the necessity of human involvement in authorship and copyright protection − though this judgment may have been different in a UK or EU court, with the photographer probably receiving the rights.

Proposals for change

A great many proposals aim to address GenAI authorship, such as:

• Economic rights only: some suggest limiting GenAI works to economic rights, excluding moral rights. This approach acknowledges the economic value of GenAI works, without granting full authorship rights.
• Joint authorship: a model where both the AI and the human user are considered co-authors based on their contributions. This approach recognizes the collaborative nature of GenAI works.
• Partial legal capacity: inspired by German law, this approach could grant AI a limited set of rights and obligations. This would allow AI to be recognized as a creator without granting it full legal personhood.

The presence of GenAI works in the arts and their implications for copyright law demand a nuanced understanding of the law on originality and authorship and an appreciation for the human element of art. While current frameworks offer a foundation, evolving legal interpretations and legislative reforms are essential to keep pace with technological advancements. As AI advances, frameworks must adapt − this includes considering the unique nature of GenAI works and finding ways to balance the interests of human creators with demand for innovation.

Read this Oracle to find out more about using AI in law firm applications.

Personal reflections

I found my dissertation to be a deeply enriching experience, which significantly sharpened my understanding of the legal complexities and broader societal implications of GenAI in the creative field − but by no means fully sated my curiosity!

I hope this exploration of AI and copyright law has piqued your interest in this topic. I should reiterate that this is a very brief overview of an incredibly vast area of legal research and debate. As such, many points here are brief, if not plainly reductive. Engaging with these questions is not merely an academic exercise, but a necessary step towards understanding and shaping the future of creativity and innovation. I hope that if you have an interest in this issue, you can continue to engage with this fascinating area of the law in your studies!

Keep an eye out for a future blog from me, where I’ll share some general insights and advice for writing your own dissertation.