On 24 October 2019, the University of Johannesburg (UJ) hosted the Industry 4.0 and Intellectual Property (IP) Colloquium organised by Prof Wim Alberts of UJ’s law faculty. The main objective of the colloquium was to discuss the impact of the fourth industrial revolution (4IR), especially artificial intelligence (AI), on IP. Speakers at the colloquium were drawn from academia, government and largely from practice and industry. Yours sincerely was there to register the presence of the University of Cape Town’s NRF/DST/SARChI Chair on IP, Innovation and Development.
After a warm welcome from Prof Letlhokwa Mpedi, Executive Dean of UJ’s law faculty, the colloquium commenced with a very educating and authoritative keynote address by Prof Coenraad Visser of the University of South Africa (UNISA). The keynote address did not only set the tone/agenda for discussion, but also summed up and foreshadowed the key issues addressed at the colloquium, especially in relation to the impact of AI on knowledge creation, IP regimes, and the ‘deepening tension between the explicit welfare objectives of trade policy’ and the existing global IP legal architecture that shapes transactions in the digital and information markets.
The possibilities and perils of the 4IR
Prof Visser kicked off his address by noting that the 4IR is a ‘new chapter in human development, enabled by extraordinary technological advances’ with a speed, breadth and depth that is forcing us to reconsider ‘how countries develop, how organisations create value and even what it means to be human’. According to him, the 4IR is ‘merging the physical, digital and biological worlds in ways that create both huge promise and potential peril’ to human existence.
The possible perils of the 4IR, according to him, includes the generation of greater inequalities particularly in the disruption of the labour market, monumental increase in the vulnerabilities present in the digital space, and the ethical concerns that would emerge from automation, robotics and genetic engineering, among others. On the other hand, the 4IR has great promises of lowering barriers between inventors and markets due to new technologies such as 3D printing, integration of different technologies and domains, increasing internet connectivity, and improved quality of lives through the use of robotics, and a more active role for AI.
Zeroing in on the main objective of the colloquium, Prof Visser spoke specifically on the implication of AI on copyright and patent law. Here, he took note of the prediction of experts to the effect that AI technology has 50% chances of reaching human-level intelligence by 2040 and a 90% likelihood by 2075. Thus, he argued, the more sophisticated the AI technology becomes, the less human intervention would be required in creating copyright works and generating patentable inventions.
4IR and fair use
A very important aspect of Prof Visser’s keynote is its recognition of the potential of AI technology to threaten the ‘fair use doctrine from a variety of technical and private law mechanisms, including technical protection measures and anti-circumvention law. The importance of this aspect is based on the fact that the fair use doctrine, which is contained in the proposed section 12B in clause 13 of the Copyright Amendment Bill [CAB], formed the subject of condemnation by Stephen Hollis in his presentation on the ‘real reasons behind Government’s drive to enact the [CAB]’.
Arguing that government formulated the CAB mainly to serve the interest of local lobbying groups backed by big international tech companies, Hollis maintained that the fair use (anti-circumvention and other) exceptions being proposed should be regarded as “free use” because, according to him, it will enable all uses of copyright works without paying royalties to copyright owners. Interestingly, Hollis believed that the CAB will be an effective legal regime for the 4IR. He was also not against the educational exceptions in the CAB. His major concern, however, is the fair use exception, which he believed went overboard in favour of “users”, whom he claimed is government and big tech. Hollis further argued that no other country in the world has a fair use provision, even though he mentioned at the same time that the provision in the CAB was inspired by the US fair use exception. He believed to be a lie the fact that fair use exception is contained in the copyright legislations of countries such as Liberia (section 9.8) and Philippines (section 185), among others.
It is amazing how Hollis acknowledged the usefulness of the CAB within the 4IR and in the same breadth condemned its innovative provisions, which will not only position South Africa to benefit from the gains of the 4IR, especially the AI technology (as eloquently argued by Dr Andrew Rens here), but will also engender social justice within the copyright based industry and the larger economy. Indeed, it appears that Hollis did not pay attention to the golden words from Prof Visser’s Keynote on the positive impact of fair use in the digital age as follows:
‘Fair use is an open-ended exception that facilitates some unauthorised uses of copyright works. Rooted in English common law, the doctrine has enabled a wide range of interactions on the internet, from cultural and political engagement to economic transactions. Many of the permission-less innovations for which the digital age is known emerged from the tapestry of a rich fair use landscape. Image search engines, linking, and device interoperability, for example, all exist in the large part because of limitations and exceptions to copyright, typified by a robust fair use/limitations and exceptions environment. In short, consumer driven entitlements are an indispensable component of the digital economy’.
AI Author and copyright law
Prof Visser noted AI’s potential to transform and stretch the boundaries of three key concepts. First, he pointed out that given its capacity to independently create copyright works, AI continues to challenge the bedrock principle of authorship (and related originality) which is currently ‘based on assumptions about human behaviours, judgment and activity’. ‘Though the authorship debate engenders controversial notions about the romantic and philosophical source of creativity’, Prof Visser argued that ‘copyright has been, and continues to be, hostile to nonhuman [and non-juristic] authors’. The point has earlier been echoed in our (Prof Caroline Ncube and I) paper on the subject of non-human authors. Nonetheless, John Foster further elaborated on the issue in is presentation at the colloquium.
Foster, who spoke on ‘AI and the copyright conundrum’, agreed that the Copyright Act does not currently recognise the non-human or non-juristic author. However, relying on judgments of the courts works created with little or no human intervention (computer-generated works), he argued that some protection could be afforded AI generated works if treated as computer-generated works, in which case authorship (and ownership) will be conferred on the person who made the arrangement for the work (such person would be the AI developer). However, as reported in an earlier blog, the Beijing Internet Court refused to grant copyright over a work exclusively created by AI.
Another impact of AI in the copyright legal domain, according to Prof Visser, is the diminution of the first sale or exhaustion doctrine especially within the context of digital trade. According to him, as a consequence of AI, ‘licensing, rather than outright transfer of ownership, has emerged as a dominant paradigm for trading in digital goods like entertainment media and software’. This development, he argued, ‘threatens to eclipse the economic rationale underlying exhaustion and the entities and markets that depend on it by eliminating the ability for good faith end users to benefit from a secondary market of any kind’.
AI Inventor and patent law
According to Prof Visser, AI raises four questions within the context of patent law, namely: (a) the patent subject-matter eligibility of AI technologies; (b) AI’s role in the definition of “person of ordinary skill in the art” in the non-obviousness standard; (c) the inventorship of AI-generated inventions; and (d) liability for patent infringement by AI. These issues were examined in more depth by speakers, such as Madelein Kleyn and John McKnight.
In their respective presentations, Kleyn and McKnight recognised the fact the Patent Act does not envisage a non-human inventor such as AI. However, while Kleyn argued that it is not very clear-cut if the provision of the Patent Act defining patentee excludes AI, McKnight warned against applying the Patent Act to recognise a third juristic person (electronic personality) by conferring patent on AIs, no matter how independent and advance they become. In this connection, he preferred the approach being adopted by the European Patent Office (EPO) on an application for patent in respect of a food container that uses fractal designs. The application lists DABUS (an AI) as a co-inventor with its developer, Stephen Thaler. DABUS is a creative machine which is developed to generate ideas without human intervention. Regarding DABUS as a mere inventing tool, the EPO has indicated an unwillingness to recognise it as a co-inventor.