Can Artificial Intelligence be an Inventor under Patent Law? Australian Federal Court says ‘Yes’ but U.S. District Judge says ‘No’

On 18 August 2021 we published a commentary on Thaler v Commissioner of Patents [2021] FC 879 (“the Australian Case”) where the Federal Court of Australia held that an artificial intelligence (“AI”) system can be named as an inventor in a patent application under the Patents Act 1990 (“Australian Patents Act”).  Our commentary on the Australian Case can be accessed here.
 
On 2 September 2021, United States District Judge Leonie Brinkema issued a memorandum opinion in Stephen Thaler v Andrew Hirshfeld, et al (Case 1:20-cv-09 (LMB/TCB)) (“the US Case”) wherein she opined that an AI system cannot be an inventor under the United States’ Patent Act (35 USC) (“US Patent Act”).
 
We will now discuss the decision in the US Case.
 
Background
 
The US Case involved the same Plaintiff and AI as the Australian Case i.e. Dr Stephen Thaler and the AI known as DABUS. Dr Thaler similarly made an application with the United States Patent and Trademark Office (“USPTO”) to patent two inventions: a light beacon that flashes in a new and inventive manner to attract attention and a beverage container based on fractal geometry.
 
DABUS is identified as the inventor in the patent applications. The US Patent Act requires an inventor to execute an oath or declaration, and because DABUS could not do so, Dr Thaler included in his application a “Substitute Statement” which explained that DABUS was “under legal incapacity” since it is an AI with no legal personality or capability to execute an oath.
 
The USPTO did not accept that DABUS can be named as an inventor under the US Patent Act. Dr Thaler subsequently filed an administrative review action to challenge the USPTO’s decision in the US Case.
 
The Decision in the US Case
 
The District Court found that an artificial intelligence machine cannot be an inventor based on the plain statutory language of the US Patent Act and did not accept Dr Thaler’s arguments that policy considerations should override the plain statutory language.
 
The definition of an “inventor” in the US Patent Act makes specific reference to the word “individual” as follows: “each individual who is the inventor or a joint inventor execute an oath or declaration…”. Thus, the District Court proceeded to interpret the meaning of the word “individual” and concluded that the plain meaning of an “individual” is “a natural person”. The District Court also referred to Univ. of Utah v Max-Planck Gesellschaft 734 F.3d 1315, 1323 and Beech Aircraft Corp v Edo Corp 990 F.2d 1237, 1248 where the Federal Circuit had ruled that inventors must be natural persons.
 
Dr Thaler attempted to persuade the District Court that the US Patent Act’s reference to inventors as “individual” was not made in the context of AI-generated inventions and that policy considerations such as encouraging innovation should be taken into account when interpreting the US Patent Act. The District Court was not persuaded and took the position that matters of policy are for Congress and could not overcome a statute’s plain language.
 
Commentary
 
The reason for the difference between the decision in the Australian Case and the US Case is the difference in language between the patent statutes of Australia and the United States. The US Patent Act contains plain statutory language requiring an inventor to be a natural person, whereas the Australian Patents Act does not, and that allowed Justice Beach to venture further into the exercise of purposive statutory interpretation in the Australian Case.
 
As set out in our commentary on the Australian Case, the Malaysian Patents Act appears to be more like the US Patent Act in that the statute contains clear language which envisions the inventor to be a human capable of signing a declaration in writing.
 
To sum up, it would appear that the question of whether an AI can be an inventor under patent law will depend on the specific language in the patent statute of each jurisdiction. In jurisdictions where there is plain statutory language requiring an inventor to be a natural person, it is likely that legislative amendments will be required to explicitly incorporate and recognise AI generated inventions. On the other hand, in jurisdictions such as Australia where the statutory language is less clear, the courts may have more leeway to consider purposive statutory interpretation and policy considerations.
 
It has been reported that Dr Thaler will appeal the District Court’s decision.1
 
Article by Lam Rui Rong (Associate) of the Intellectual Property Practice of Skrine.
 

1 US judge rejects bid for patent by AI ‘inventor’, The Sunday Star, 5 September 2021.

This alert contains general information only. It does not constitute legal advice nor an expression of legal opinion and should not be relied upon as such. For further information, kindly contact skrine@skrine.com.