In this era of the Fourth Industrial Revolution, advances in artificial intelligence ("AI") has resulted in AI capable of generating inventions that are novel and inventive. The question becomes whether these AI generated inventions can be protected under the current patent law framework.

A recent development in Australian jurisprudence takes a step toward clarifying the applicability of patent law on AI generated inventions. The Federal Court of Australia recently held in Thaler v Commissioner of Patents  [2012] FC 879 ("the Thaler case") that an AI system can be named as an inventor in a patent application.

After a brief introduction on basic concepts underlying patent law and AI, this article will discuss the Thaler case followed by an analysis on the Malaysian perspective.

Difference between an owner and inventor of a patent

In a patent application, the owner and inventor can be different entities. The person making an application to patent an invention is known as the "applicant". If the application is accepted, the "applicant" becomes the "owner" of the patent. The applicant/owner of the patent is responsible for the filing of the application and if accepted, will be granted the monopoly rights over the invention which can be enforced against anyone infringing the patent.

The inventor can choose to be the applicant/owner of the patent, but the inventor can also assign the rights to a different entity. A traditional example is where employees assign their rights over their inventions to their employers when made in the course of employment.

In the current discourse over the applicability of patent law on AI generated inventions, the consensus is that the applicant/owner of the patent must be a human. There are various reasons for this including clear wordings of patent statutes requiring an applicant/owner to be a human and the lack of true autonomy by an AI to be able to independently apply for and maintain a patent. On the other hand, there is less clarity on whether an inventor must be a human or not. This is the point of contention in the Thaler case.

What is an AI?

There is no single universally accepted definition of AI, and this article does not explore the complex technicalities underpinning an AI system. Suffice to say, an AI is a machine that learns from input and data to develop the capacity for reasoning, pattern recognition and output without human intervention. Thus an AI is, in essence, a learning system. It is different from a software program where the output is based on a set of fixed algorithms. The implication is that it is possible for an AI to generate an output independently without human intervention.  

The Thaler case

Dr Stephen Thaler submitted an application to patent an invention for products and methods concerning containers, devices and methods for attracting enhanced attention using convex and concave fractal elements. In his application, Dr Thaler named himself as the applicant of the patent and the inventor as the AI system – a device for the autonomous bootstrapping of unified sentience ("DABUS"). Dr Thaler is the owner and controller of DABUS, and also owns the copyright in DABUS's source code and the computer on which DABUS resides.

The Deputy Commissioner of Patents ("Deputy Commissioner") rejected Dr Thaler's application on the basis that DABUS, an AI, cannot be named as an inventor under the Patents Act 1990 ("Australian Patents Act"). Dr Thaler then sought judicial review of that decision to the Federal Court of Australia ("Federal Court").

At the Federal Court, Justice Beach held that an AI can be named as the inventor in the patent application and remitted Dr Thaler's application back to the Deputy Commissioner to be assessed for formalities.

At the core of Justice Beach's reasoning is a focus on the underlying objective of the patent law system to incentivize innovation and development. Justice Beach is of the view that recognising an AI system as an inventor would encourage and incentivize the development of more AI systems which can generate inventions that can lead to new scientific developments. If the law does not recognise that an AI system can be an inventor, this would create uncertainty in determining the identity of the inventor of patentable AI inventions.

Justice Beach's decision was supported by the following analysis:

  • there are no express statutory provisions in the Australian Patents Act that preclude an inventor from being a non-human AI. The word "inventor" is not defined and a broader interpretation should be taken on the concept of an inventor so as not to inhibit innovation which may benefit from the output of an AI; 
  • it is consistent with the object of the Patents Act 1990 to construe the term "inventor" in a manner that promotes technological innovation and the publication and dissemination of such innovation; 
  • there is no specific aspect of patent law that would drive a construction of the Australian Patents Act as excluding non-human inventors - there is no ground of invalidity in patent law based upon the absence of any "inventor"; and 
  • in response to the Deputy Commissioner's argument that an AI cannot be named as an inventor because an AI cannot "assign" the invention to another, Justice Beach drew an analogy to ownership of the progeny of animals or treatment of fruit or crops produced on occupied land, and held that Dr Thaler is the owner, programmer and operator of DABUS who made the invention, thus making him the owner of the invention without the need for an "assignment".

As this Federal Court decision is appealable, it remains to be seen if an appeal will be filed.

What about the Malaysian Patents Act 1986?

Foreign judgments have no binding effect on Malaysian courts. In any event, the Malaysian Patents Act 1986 ("Malaysian  Patents Act") is not identical with the Australian Patents Act. The differences in the Malaysian Patents Act are likely to result in a different finding if similar factual circumstances are brought before the courts in Malaysia. This is elaborated further below.

An "inventor" is not defined in the Malaysian Patents Act. However, it could be argued that the way the word "inventor" is used in the Malaysian Patents Act and regulations supports an interpretation that would exclude an AI system. Section 12 of the Malaysian Patents Act defines "invention" as "an idea of an inventor which permits in practice the solution to a specific problem in the field of technology". In addition, Regulation 6 of the Patents Regulation 1986 provides that:

  1. the application shall contain the name and address of the inventor; and
  2. an inventor who does not wish to be named in a patent may, by declaration in writing signed by him and submitted to the Registrar, indicate that he does not wish to be so named

In particular, the language in the Patents Regulation 1986 is quite clear that the statute envisions the inventor to be a human who is capable of signing a declaration in writing. This would therefore effectively preclude an AI system.

It is likely that the Malaysian Patents Act may need to be amended in the future to incorporate and recognise AI generated inventions.

Conclusion

While the decision in the Thaler case may not be directly applicable to the Malaysian context, the principles and analysis on the importance of protecting AI generated inventions is illuminating. AI generated inventions are a current reality and will become more common occurrences due to the rapid advancement in AI technology.

It is time for the law to acknowledge that there is value in incentivizing more AI generated inventions – whether through introduction of new laws, amendment of existing laws or judicial precedents. In the words of Justice Beach: "We are both created and create. Why cannot our own creations also create?"

Originally Published 18 August 2021

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.