Written by Thomas Staveley | June 30, 2022
This article will discuss artificial intelligence (AI) and how it intertwines with intellectual property law, and the difficulties AI presents for IP law as it becomes more autonomous. We’ll consider the authorship of works generated through AI, and question whether we require legal reform in the UK to account for autonomous machine learning.
There is no one definition for artificial intelligence. The English Oxford Living Dictionary defines AI as: “the theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speed recognition, decision-making, and translation between languages”[1]. Meanwhile, the Encyclopaedia Britannica defines AI as: “the ability of a digital computer or computer-controlled robot to perform tasks commonly associated with intelligent beings”[2].
Importantly for intellectual property law, there are several high-profile examples of AI “creating” works. In 2018, a piece of art created by means of artificial intelligence was sold at Christie’s auction in New York for approximately, $432,500. This work was called ‘Edmond de Belamy, from La Famille de Belamy’ and was created through a collaboration between a human and a machine, which used an algorithm to make certain creative choices.
The sale of the work was a milestone for artificial intelligence and the creative industry; however, this was not the first work of art to be created as a result of artificial intelligence. Specifically, in 2016, the ‘Next Rembrandt’ was created by a computer which examined several Rembrandt original works and used what it learnt to create a new, original work.
Where new works are being created, the question of copyright is raised.
And for copyright to vest in a literary, dramatic, musical or artistic work, it needs to be original. Originality is said to be where an author has created the work through his own skill, judgement and individual effort [Ascot Jockey Club Ltd v Simons][3]. The threshold for this is low. However, the focus here is in relation to authorship and originality in the context of computer-generated works through AI.
An author is defined under the Copyright, Designs and Patents Act, 1988 (the “Act”) as the person who created the works[4]. However, this is slightly different when it comes to works which are computer generated. In this instance, “the author shall be taken to be the person by whom the arrangements necessary for the creation of the work are undertaken”.
But, advancements in AI have led many to believe the current legislation is not capable of protecting works created by AI, especially where AI systems are developing through machine learning to a stage which allows the AI to be capable of self-learning. In this instance, questions arise as to authorship of AI works where an AI which is capable of making autonomous decisions can still be protected by copyright.
A case in South Africa decided that works created by a computer will enjoy copyright protection as the creator of the work will have made the relevant arrangements for the creation of the work by the computer. In this case, the defendant disagreed and claimed that an individual who creates a computer-generated work is not the author and as a result, ownership does not vest in the author. Whitford J responded rather nicely, stating, “It is as unrealistic as it would be to suggest that, if you write your work with a pen, it is the pen which is the author of the work rather than the person who drives the pen.”[5]
Nonetheless, AI is continuing to evolve and develop to allow for “machine learning”[6], which includes unsupervised and supervised learning[7]. Unsupervised learning sees AI capable of determining specific patterns within large amounts of data of which the author is not aware [8]. As a result of this lack of knowledge and awareness by the author with respect to how the AI may operate and subsequently make unsupervised decisions to create a work can lead to questions as to whether the final work is then still a work of the author’s own intellectual creation.
For example, in an exhibition named “UUmwelt”, Pierre Huyghe gathered significant amounts of MRI data which was then placed into a network[9]. This network, alongside various external factors outside of Pierre Huyghe’s control (including the specific lighting and humidity of the venue at the time of the exhibition), created the work. Pierre Huyghe acknowledged that “what is created is not necessarily due to the artist as the only operator”.
Pursuant to section 16 of the Act, copyright infringement takes place by a ‘person’[10]. However, as discussed, where AI has developed into being capable of machine learning to make decisions outside of the author’s control and subsequently infringes copyright as a result, can it still be claimed that the author has infringed copyright?
This is especially intriguing where the author gave no such instruction to the AI, and therefore posits the question of where do the arrangements necessary to create the work start and stop.
As the law stands, the author who made the necessary arrangements to create the work, whether that be initially inputting the data or something else, will be held liable.
We have discussed what constitutes an author of computer-generated works, however, as advances in artificial intelligence continue to develop, does UK law give scope to allow for these changes?
In June 2022, the UK Government published its response to its consultation on Intellectual Property and forms of Artificial Intelligence. The consultation touched on copyright protection for computer-generated works where it can be argued there is no human author (as explained above), certain exceptions to copyright for text and data mining which is significant in artificial intelligence use and patent protection for artificial intelligence devised inventions.
Following the initial consultation, the Government believed that there was no need for change with respect to computer-generated works and patents devised by artificial intelligence. However, it was accepted that certain changes need to be made with respect to text and data mining. In its recent publication with reference to this consultation, the Government has put forward its intentions to revise current UK copyright law to make it more accessible to review and analyse data for the purposes of machine learning, research and innovation.
With respect to patents devised by AI and computer-generated works, the Government has stuck with its decision that there is no need to make any reforms to UK patent and copyright law as it stands today, as it believes the current legislation is sufficient. Nevertheless, the Government has stated it will continue to keep the developments of artificial intelligence and machine learning under review to ensure that rules governing inventorship and authorship continue to coincide with the developments of artificial intelligence innovation.
If you require further information or help with AI and IP law, please contact info@briffa.com to arrange a free consultation with one of our expert lawyers.
[1] Lexico, British & World English: Artificial Intelligence’ (Lexico, 2019)
[2] B.J. Copeland, ‘Artificial Intelligence’ (Encyclopaedia Britannica 2019)
[3] Ascot Jockey Club Ltd v Simons [1968] 64 WWR 411
[4] S.9(1) Copyright, Designs and Patents Act 1988
[5] Payen Components South Africa Ltd v Bovic Gaskets CC and Others (448/93) [1995] ZASCA 57
[6] David Goldberg and John Holland, ‘Genetic Algorithms and Machine Learning’ (1988) 3 (2-3) Machine Learning
[7] Timothy Pinto, ‘Robo ART! The copyright implications of artificial intelligence generated art’ (2019) 30 (6) E.L.R
[8] Ibid.
[9] Simon Ings, ‘Pierre Huyghe at the Serpentine – digital canvases and mind reading machines’ (Financial Times 04 October 2018)
[10]Copyright, Designs and Patents Act 1988, s 16 (2)
Reaffirming the principle that AI inventions must meet the same patentability criteria as other computer-implemented inventions
The recent Court of Appeal decision in EWCA Civ 825 has significant implications for the patentability of AI inventions in the UK. The Court of Appeal reversed the High Court’s…
Sexualised & Synthetic: The Legal Implications Of The Celebrities Most Commonly Victimised By Deepfakes
‘Taylor Swift deepfake’ is searched 9,900 times per month – the most of any celebrity. ‘Deepfake Pornography’ has increased in searches by 97.5% since 2019. 25% of the most searched…
We’ll start with a no obligation chat where we’ll get to know you and understand your current challenges.
Contact us now