The Question of Authorship in Computer-Generated Work
Nahide Basri is a Cypriot LLB graduate at the London School of Economics and Political Science (LSE); she is currently undertaking a Master of Laws specializing in Human Rights Law. She has received the Morris Finer Prize for the Best Performance in Family Law and Hunton Andrews Prize for the Best Performance in information Technology and the Law during her LLB at LSE.
The Question of Authorship in Computer-Generated Work
The increasing involvement of Artificial Intelligence (broadly – ‘computers’) in the creative process is challenging our traditional assumptions about authorship in cases of Computer-Generated Work (Grimmelmann, 2016). Contrary to the beliefs of many scholars, including the previously cited Grielmannn, this extract argues that Computer-Generated Work should not be forced into the existing copyrights framework. Instead, Computer-Generated Work-specific legal framework is necessary, Section 9(3) of the Designs and Patent Act (CPDA) 1998 presenting a good model in this direction (Bridy, 2012).
Traditionally, where computers are utilized to generate a predicted outcome under the direct guidance of human authors (e.g. writing a poem on Microsoft Word), their use as a mere tool poses no challenge to identifying the creator of the work (Grimmelmann, 2016). However, where computers act as independent actors, generating Computer-Generated Works algorithmically, sequentially, or non-deterministically, there is an apparent gap between the human’s input and the computer’s output, threatening the “individual creative personality” of the monolithic romantic author model (Bridy, 2012). Determination of the author gets even more complicated when Computer-Generated Works are created by a computer program that is coded and used by two distinct individuals (Grimmelmann, 2016). There are now at least three different potential authors: the computer-programmer, the computer-user, and the computer itself.
Even then, it is not immediately clear why the traditional copyrights regime cannot accommodate Computer-generated Work. On the one hand, scholars correctly argue that the aforementioned challenges are not completely unprecedented (ibid). Arguably, there are plausible arguments against granting computers copyrights. For instance, even where computers appear like independent artists, Computer-Generated Work is (currently) produced under the programmers’ initial instructions and parameters (Murray, 2018), bearing a “sufficient nexus to human creativity [of the programmer]” (Bridy, 2012). Furthermore, granting copyrights to computers may also be undesirable in policy terms, since computers, unlike humans, need neither their ‘rights’ to be recognized, nor economic incentives to be creative (Hristov, 2017; Cohen, 2007). Thus, under this view, it seems that the challenge boils down to allocating copyright ownership between the user and the programmer (Grimmelmann, 2016). The digital dimension of this choice should not blind us to the fact that similar decisions are already made offline where various actors contribute to the final work (ibid). In both online and offline contexts, judges engage in a subjective analysis of each contributor’s input of skills (e.g. in Nova Productions Ltd v Mazooma Games), originality (s.1(1) CPDA), and creativity (e.g. in Infopaq International) (Grimmelmann, 2016).
As tempting as such internal perspectives equating Computer-Generated Works to human creativity are, the more the computer’s output exceeds the users’ and coders’ input, the more difficult it gets to call either an author in the traditional sense (Bridy, 2012). A striking example is Google’s Deep Dream which creates unique images via an algorithm – which is less clear and precise than a simple algorithmic creation (Guadamuz, 2017). It thus exemplifies a wider trend of the unprecedented ability of computers to mimic human creativity by displaying similar traits such as originality and unpredictability (Bridy, 2012). In this context, claiming that Computer-Generated Work is conceptually identical to “old-fashioned pen-and-paper works” (Grimmelmann, 2016) is to ignore the pluralistic creation of the Computer-Generated Works – resulting from the contributions of the human author and the computer (Cohen, 2007).
This therefore renders Computer-Generated Work-specific legislation necessary. It is argued that Section 9(3) is a good model in this direction. By rendering the “person…[who made]… the arrangements necessary for the creation of [Computer-Generated Works]” the author, Section 9(3) clearly acknowledges the legal fiction involved in vesting copyright in the human programmer/user, the author-in-law, who might not be the (only) author-in-fact (Bridy, 2012). This demonstrates a conscious policy choice to vest the copyright in the former despite the fact that he/she might not truly be the (sole) ‘author’ of the work in question. Furthermore, despite scholars’ concerns that Computer-Generated Work-specific legislation will “blind us to the immense diversity that category [Computer-Generated Works] embodies,” (Grimmelmann, 2016) the existing case-law on Section 9(3) depicts a case-by-case approach. For example, in Express Newspapers, preceding Section 9(3), where the computer was used like a “pen” – and a “tool” – the copyright was granted to the user. Alternatively, in Nova Productions, where the court had to determine whether the graphic work in question was the creation of the programmer or the user of the program, it chose to grant the copyright to the former. Specifically, Judge Jacob LJ evaluated the potential authorship of the user, eventually rejecting it on the grounds that his input is “not artistic in nature and nor has he contributed to skill or labour.” This depicts that Section 9(3) permits a flexible framework which takes into account the different ways in which Computer-Generated Works can be generated.
Undoubtedly, scholars such as Grimmelmann, are correct in objecting to the much exaggerated gap between computer and human creativity (Grimmelmann, 2016). Despite this, the element of artificiality involved in collapsing the (widening) gap between the author-in-fact and author–in-law should be conceded through Computer-Generated Work-specific legislation (Bridy, 2012). To this end, Section 9(3) appears to be the ideal solution, incorporating Computer-Generated Works within the copyrights framework by acknowledging that computer creativity is “something other than (but owing to) the human creativity of its coder.” (ibid).
- Grimmelmann, J. (2016). There’s No Such Thing as a Computer-Authored Work – And It’s A Good Thing, Too. Columbia Journal of Law & Arts, 39: pp.403-416
- Bridy, A. (2012). Coding Creativity: Copyright and the Artificially Intelligent Author. Stanford Technology Law Review, 5: pp. 1-28
- Hristov, K. (2017). Artificial Intelligence and the Copyright Dilemma, The Journal of the Franklin Pierce Centre for Intellectual Property, 57: pp. 431-454
- Cohen, J. E. (2007). Creativity and Culture in Copyright Theory, the UC Davis Law Review, 40: pp 1151-1205
- Nova Productions Ltd v Mazooma Game (2007) EWCA Civ 219
- Express Newspapers v Liverpool Daily Post & Echo (1985) 1 WLR 1089
- Infopaq International A/S v Danske Dagblades Forening (2009) ECDR 16 (Case C-5/08)
- Copyright Designs and Patents Act 1988
- Guadamuz, A. (2017. Do Androids Dream of Electric Copyright? Comparative analysis of originality in artificial intelligence generated works, Intellectual Property Quarterly