 |
| David Knight |
It is important to note that the judgment, in the case of Stephen Thaler v. Comptroller-General of Patents, Designs and Trade Marks,[1] does not address the question of whether AI-generated inventions meet the requirements for patentability, but considers instead questions around inventorship and ownership.
The patent applications in question relate to:
- A food or beverage container in which the walls of the container have a fractal profile to allegedly improve grip and enable interlocking of containers; and
- A light beacon that flashes in a particular pattern and frequency, which allegedly is more catching to the human eye and therefore more likely to be spotted in an emergency.
Stephen Thaler was the driving force behind the patent applications. Thaler is the president and CEO of Imagination Engines, and the owner and creator of DABUS, the AI system that created the inventions.
Thaler might have, but did not, name himself as an inventor, i.e., the actual deviser of the invention and, therefore, applicant for the patent applications.[2] Instead, Thaler contended throughout that the DABUS AI system conceived the inventions, i.e., DABUS is, and should be named as, the inventor, and that he was entitled to own the inventions because he owned DABUS. As the judgment records:
[I]n this jurisdiction, it is not and has never been Dr. Thaler's case that he was the inventor and used DABUS as a highly sophisticated tool. Had he done so, the outcome of these proceedings might well have been different.[3]
This article considers the rationale for the Supreme Court's decision that Thaler's patent applications failed, alternative approaches to seeking patent protection for AI-generated inventions that might avoid the same outcome, and some thoughts on how the decision might affect infringement and validity disputes around such patents.
The United Kingdom Patents Act 1977 sets out who has the right to apply for and obtain a patent. The starting or default position in Section 7(2)(a) of the Patents Act is that a patent may be granted to the inventor, but, according to Sections 7(2)(b) and (c), this default position can be changed where the inventor has contracted with someone — typically an employer — or where there are successors in title.
Section 7(2) finishes with the words "and to no other person." One needs therefore to consider who is the inventor, i.e., the devisor of the invention, and everything then flows from that.
After considering the references "persons" and "inventor" in the Patents Act, the court concluded that there is:
[O]nly one interpretation: an inventor within the meaning of the 1977 Act must be a natural person, and DABUS is not a person at all, let alone a natural person: it is a machine and on the factual assumption underpinning these proceedings, created or generated the technical advances disclosed in the applications on its own.[4]
As there is no inventor, it necessarily follows that there was no starting point for determining to whom a patent could be granted. The Supreme Court also held that, even if that hurdle had been overcome, merely being the owner of DABUS was not sufficient to transfer ownership of the rights in the invention to Thaler.
U.K. law requires that the applicant for a patent file a statement of inventorship naming the inventor or inventors.[5] As Thaler did not do this nor remedy the defect within the statutory time frames, his patent applications failed, which was further compounded by his failure to explain how he had the right to file the patent applications.
The case has always been seen as a test case taking on the sections in the U.K. patent legislation around inventorship. The judgment does not address the question about what would have happened had Thaler kept quiet about the involvement of the DABUS AI system in making the inventions and had just filed the patent applications naming himself as inventor and applicant.
The U.K. Intellectual Property Office likely would not have looked behind such statements and would have allowed the patent applications to proceed. Had they subsequently been granted, it then would have been up to third parties to challenge those patents on the basis of lack of human inventors.
Such a challenge could be along the following lines. The invention was made by an AI system, and therefore there is no inventor. If there is no inventor, then there is no individual from whom the patent applicant can derive title. Accordingly, the applicant is not entitled to the invention.
In the U.K., this would then provide the basis of a defense to a claim of patent infringement on the basis that "third parties [in defense of a patent infringement claim] can challenge title to sue, just as a man sued for trespass can challenge the claimant's possession of the land over which he is said to have trespassed."[6]
It would not, however, provide a basis for seeking revocation of the patent as, under U.K. law, only a person otherwise entitled to the patent can seek its revocation on the ground that it was granted to someone not entitled to it.[7]
Since there is no inventor in this case, and therefore no originating chain of title, there cannot be anyone with a claim to a superior title which would enable a claim for revocation on the basis of lack of entitlement to get off the ground.
In the years to come, we may see cases addressing these issues. In the meantime, the sensible approach for a Thaler-type patent applicant would be to position the AI as a tool that is used by an individual in making the invention, for that person's name to go on the statement of inventorship and to get away from the Thaler approach of naming the AI as the inventor.
Given that it will often be the case that there is some human oversight or direction of the AI, it is not a stretch to describe the AI as a tool. But, if the invention is truly created by a completely autonomous AI system, then U.K. law is that it will not be patentable as there is no human inventor.
Even if the AI-as-tool route is used to secure patent protection, that might give rise to some interesting questions in due course. For example, who is the notional skilled person? A human inventor or a human and AI team? And allied with this, is whether there is a different standard of test around obviousness for AI-conceived inventions compared with human-conceived inventions.
One school of thought is that if AI just pops out the solution, then the solution must be obvious, and therefore not patentable; but on the other hand, if the solution has evaded human inventors, then it suggests that the solution is inventive and not obvious, and therefore patentable. Interesting times!
In the meantime, it is now, however, settled law in the U.K. that an AI system cannot be an inventor — at all times it is necessary for the named inventor to be a natural person, and rights of ownership will flow from whoever that human inventor is, or inventors are.
David Knight is a partner at Brown Rudnick LLP.
The opinions expressed are those of the author(s) and do not necessarily reflect the views of their employer, its clients, or Portfolio Media Inc., or any of its or their respective affiliates. This article is for general information purposes and is not intended to be and should not be taken as legal advice.
[1] I.e., the UK Intellectual Property Office.
[2] Section 7(3) of the Patents Act.
[3] Paragraph 52 of the Supreme Court Judgment.
[4] Paragraph 56 of the Supreme Court Judgment.
[5] Section 13 of the Patents Act and rule 10(3) of the Patent Rules 2007 (SI 2007/3291)
[6] Cinpres Gas Injection Limited v. Mela Limited
[2008] RPC 17, [2008] EWCA Civ 9 at 117.[7] Sections 72(1)(b) and 72(2) of the Patents Act.
For a reprint of this article, please contact reprints@law360.com.
