In February 2024 the United States Patent and Trademark Office (USPTO) issued guidance setting out principles for determining whether or not an AI-assisted invention can be deemed to have a human inventor. In this article we give an overview of the guiding principles set out by the USPTO, and briefly compare them with the current approach taken by the UK Intellectual Property Office (UKIPO) for assessing the patentability of AI-assisted inventions.

Purpose of the guidance

Kathi Vidal, the Under Secretary of Commerce for Intellectual Property and Director of the USPTO, has stated that the US patent system was developed “to incentivise and protect human ingenuity and the investments needed to translate that ingenuity into marketable products and solutions”. Therefore, to assess the patentability of an AI-assisted invention, it is necessary to consider the extent of human contribution towards the invention, and determine whether or not that contribution was great enough to warrant the reward of patent protection.

Accordingly, the USPTO has determined that for an AI-assisted invention to be patentable the human contribution must be “significant”, so that the human is considered an inventor: similarly to what each of a group of persons would need to contribute to be considered joint inventors. The USPTO has also set out five guiding principles that provide useful context for what the USPTO may in practice consider a ‘significant’ contribution.

USPTO guiding principle 1: “A natural person’s use of an AI system in creating an AI-assisted invention does not negate the person’s contributions as an inventor”.

This guiding principle reflects the general principle that an invention is not inherently unpatentable if an AI system has been used in its creation.

The mere use of AI by a natural person in the invention creation process does not preclude that person from being an inventor, provided that they have contributed significantly to the creation of the invention.

This principle is akin to that decided in Hobbs v United States, Atomic Energy Comm (451 F.2d 849, 864 (5th Cir. 1971))), and quoted in Shatterproof Glass Corp v Libbey-Owens Ford Co (758 F.2d 613, 624 (Fed. Cir. 1985)), that “an inventor may use the services, ideas, and aid of others in the process of perfecting his invention without losing his right to a patent”.

USPTO guiding principle 2: “Merely recognizing a problem or having a general goal or research plan to pursue does not rise to the level of conception”.

A natural person who simply presents a problem to an AI system would likely not be considered an inventor of an invention output by the system as a solution to the presented problem.

However, the natural person can make a “significant” contribution by constructing the prompt provided to the AI system in view of a specific problem to elicit a particular solution. Therefore, prompt-engineering may in some cases be sufficient to meet the significant contribution criterion (see Burroughs Wellcome, 40 F.3d at; see also Hitzeman, 243 F.3d 1345, 1357-58; In re Verhoef, 888 F.3d 1362, 1366 (Fed. Cir. 2018).

USPTO guiding principle 3: “Reducing an invention to practice alone is not a significant contribution that rises to the level of inventorship”.

It is not sufficient for a natural person to merely recognise and appreciate that the output of an AI system is an invention for the natural person to be considered an inventor.

This is particularly the case when the use and output of the AI system would be readily apparent to a person skilled in the art. However, if the natural person were to take the output of the AI system and provide a significant contribution in producing an invention, then they may be considered an inventor.

Further, in some circumstances it may be possible to show that a person who conducts a successful experiment using an AI system’s output provided a significant contribution, even if the person is unable to establish conception until the invention is reduced to practice.

For example, as was the case in Dana-Farber Cancer Inst Inc v Ono Pharm Co (2019), a researcher, Dr Freeman, was found to be a joint inventor of patents claiming methods of treating cancer by administering antibodies targeting specific receptor-ligand interactions on T-cells.

Dr Freeman had used an automated search tool to identify a ligand amino-acid sequence, and performed subsequent experiments which identified several types of tumours that expressed the sequence identified by the search tool. This was a key discovery in the inventions of the subsequent patents (see Dana-Farber Cancer Inst Inc v Ono Pharm Co (2019),964 F.3d 1365, 1373-74 (Fed. Cir. 2020)).

USPTO guiding principle 4: “A natural person who develops an essential building block from which the claimed invention is derived may be considered to have provided a significant contribution to the conception of the claimed invention even though the person was not present for or a participant in each activity that led to the conception of the claimed invention”.

It is possible that, in some circumstances, a natural person who designs, builds or trains an AI system specifically to obtain a particular solution to a specific problem may be considered an inventor.

In such circumstances the designing, building and training of the AI system must be a significant contribution to the invention created with the AI system (see Dana-Farber, 964 F.3d at 1372-74).

USPTO guiding principle 5: “Maintaining “intellectual domination” over an AI system does not, on its own, make a person an inventor of any inventions created through the use of the AI system”.

It is not enough for a natural person to simply own or oversee an AI system that was used to create an invention. In order to be deemed an inventor, the natural person must be able to show that they provided a significant contribution to the conception of the invention.

This principle is also akin to that of In re Verhoef (888 F.3d 1362 (Fed. Cir. 2018)), in which it was emphasised that the person who conceives of the invention is the inventor.

AI-assisted inventions in the UK

As in the US, the UK Supreme Court has confirmed that under UK law only a natural person can be named as an inventor on a patent application (see related articles).

The UK Supreme Court noted that pure ownership of a machine does not confer to the owner the right to obtain a patent for a product or process generated autonomously by the machine, in line with the USPTO’s fifth guiding principle.

The UK Supreme Court also acknowledged that if the owner (Dr Thaler) of the AI machine in question (DABUS) were the inventor and used DABUS as a tool to generate the inventions, the outcome of the proceedings would have been different. The UK Supreme Court’s view therefore appears to be aligned with the USPTO’s first guiding principle: that mere use of an AI machine does not preclude grant of a patent.

The UK Supreme Court did not provide detailed guidance on patentability of AI-assisted inventions, in contrast to the USPTO, as questions of patentability were not the subject of the appeal. However, the position on inventorship in the UK appears to be similar to that in the US: as long as a human inventor “actually devised” the invention set out in the patent application, they can be named the inventor of an AI-assisted invention and a patent can be obtained.

In short

For both the US and the UK patent protection remains available for AI-assisted inventions provided that a human has sufficiently contributed to devising the invention. As a final practice point, documentation of the human involvement in AI-assisted inventions should be kept in order to be able to demonstrate the requisite level of the human’s contribution to the invention if required (for example, if inventorship is challenged in litigation).

Useful links

• Inventorship Guidance for AI-Assisted Inventions, USPTO, 13 February 2024: dycip.com/uspto-ai-inventions-guidance
• Stephen Thaler v Katherine K Vidal, 21-2347, US Court of Appeals Federal Circuit, 05 August 2022 (PDF): dycip.com/thaler-vidal
• Hobbs v United States, Atomic Energy Comm, 451 F.2d 849, US Court of Appeals, Fifth Circuit, 20 December 1971: dycip.com/hobbs-us-atomic-energy
• Shatterproof Glass Corp v Libbey-Owens Ford Co, 758 F.2d 613, US Court of Appeals Federal Circuit, 29 March 1985: dycip.com/shatterproof-glass-libbey-owens-ford
• Burroughs Wellcome Co v Barr Labs Inc, 40 F.3d 1223, US Court of Appeals Federal Circuit, 15 December 1994: dycip.com/burroughs-wellcome-barr-labs
• Hitzeman v Rutter, 243 F.3d 1345, US Court of Appeals, Federal Circuit, 18 May 2001: dycip.com/hitzeman-rutter
• In re Verhoef, 888 F.3d 1362, US Court of Appeals for the Federal Circuit, 03 May 2018: dycip.com/in-re-verhoef
• Dana-Farber Cancer Inst Inc v Ono Pharm Co, 964 F.3d 1365, US Court of Appeals for the Federal Circuit, 14 July 2020: dycip.com/dana-farber-cancer-ono-pharm