In July this year, the Federal Court of Australia handed down a decision in Stephen L. Thaler [2021] APO 5, which allowed listing AI system DABUS as an inventor in a patent application. It is interesting to explore what implications this decision could have in the field of copyright.
About the DABUS decision
The DABUS case refers to an international patent application where AI DABUS was listed as an inventor. Initially, it was rejected by a number of IP offices around the world (including the US, UK, EPO, German and Australian patent offices). The Federal Court of Australia heard the appeal in the case and found that under the Australian Patent Act the AI could be listed as an inventor. According to the court, Dr Thaler, who is the developer, owner and controller of DABUS, would be the owner of the patent. The court found that this is compatible with the goal of the Patent Act to promote innovation and that nothing in the Patent Act explicitly or impliedly prohibits listing AI as an inventor. The decision has been appealed to the Full Federal Court and the outcome is pending.
Implications on copyright
One could speculate about the impact this decision could have in the field of copyright law. Could this mean that courts will also gradually reconsider the AI authorship question and recognize AI as an author of AI-generated works, while the owner of AI will be recognized as the owner of these works? Would this be an appropriate solution in the field of copyright law?
It is worth noting that artists who employ AI technology in their creative practices often see AI as an important contributor in their creative processes and are often willing and ready to indicate AI as a co-author or a single author of the work. For instance, when Uncanny Valley used an AI to develop a song ‘What a Beautiful World’ (which eventually won the Eurovision AI Song Contest 2020), they wanted to list AI as one of the contributors. However, they found that current copyright law would not recognize AI as an author.
Some arguments in favor of this solution in copyright law
There are a few reasons why such emerging attribution practices could be translated into legal rules. First, as the team behind the DABUS cases argues, it might sound fair to credit actual creators, be they human or non-human. If AI autonomously, i.e. without creative contribution from a human, generates a work, then arguably it would be unfair to list a human person as an author. Second, if we agree that a human remains the owner of AI-generated works, then human authors are still the ones rewarded under the copyright law system. Also, they will be the ones responsible for commercializing the works and making them accessible to the public.
What are the problems with this approach?
At the same time, a number of arguments speak against the recognition of AI authorship. First of all, there would be some problems with fitting this new interpretation within current copyright laws. Under most copyright laws, ‘author’ is quite clearly a natural person. Even the Federal Court of Australia thought that, in contrast to ‘inventor’, ‘author’ refers to a human being. Furthermore, if AI were an author and the duration of protection is based on the author’s life, what would be the duration of protection for such AI-generated works? Even if we agree that human beings will own works generated by AI, which human (AI developer, AI owner, AI user, etc) will own them and on what legal basis? These and other questions would need to be addressed through a legal reform, since finding answers to these questions under current laws might be impossible.
In addition to the problems with current copyright law, a more significant question is whether AI is actually capable of independent creation. Some commentators disagree with the argument made in the DABUS case. They suggest that a human role is still essential in the creative or inventive process and thus the human should be named as an author or listed as an inventor. AI is just another tool, even if a more sophisticated one than previous ones, and there is no need to award a tool with an authorship.
Finally, an ethical question needs to be asked: if we provide AI with a status of ‘author’ or ‘inventor’, do we start to gradually award AI certain ‘rights’ and provide it with a certain legal status? Adopting such an anthropomorphic approach to AI might be too early and even dangerous.
Overall, while the solution that the Federal Court of Australia adopted in the DABUS case – i.e. the recognition of AI as an inventor (though not an owner) of AI-generated inventions – might be of interest when debating AI authorship questions, adoption of a similar solution in the copyright law space would not be without significant problems.
_____________________________
To make sure you do not miss out on regular updates from the Kluwer Copyright Blog, please subscribe here.
Kluwer IP Law
The 2022 Future Ready Lawyer survey showed that 79% of lawyers think that the importance of legal technology will increase for next year. With Kluwer IP Law you can navigate the increasingly global practice of IP law with specialized, local and cross-border information and tools from every preferred location. Are you, as an IP professional, ready for the future?
Learn how Kluwer IP Law can support you.
I believe the Federal Court erred, in much the same way as Birss LJ did in his minority opinion in the UK Court of Appeal decision (Thaler v Comptroller [2021] EWCA Civ 1374) in trying to interpret the current law in a way which accommodates ‘AI’, when clearly both Australian and UK law was drafted at a time when AI was neither envisaged nor considered. The courts should have remitted the problem to their respective legislatures to resolve, not least because of the fundamental issues involved. This was not just a case of embracing a technological advance within an existing framework.
The problem is that the state of the art today is not even close to true AI, which would be better understood as standing for autonomous intelligence rather than artificial intelleigence. In other words today’s so-called AI is entirely the product of human intervention and control, so it is reasonable to conclude that a human can be the beneficiary of the works so created. Indeed in respect of copyright, UK law already recognies this situation in section 9(3) of the CDPA. In my view the error lies in attributing the label of creator to the machine at a far too early stage in the evolution of AI.
Even if the Federal Court’s decision is allowed to stand, this should not be taken as a precedent for copyright or any other IPR if only for the very good reason that while patents have a fixed duration which is geared to industrial exploitation, copyright duration is fixed on the lifetime of the author (except in the case of the USA’s section 302(c) – works made for hire); this is clearly inconpatible with the idea of designating an AI device as a creator.