USPTO call for comments: Impact of AI on patentability

The USPTO has issued a request for comments regarding the impact of AI on patentability. The USPTO specifically calls for views on how the proliferation of AI could affect evaluations of patentability, including what qualifies as prior art and the capabilities of the person skilled in the art. The full call for views can be viewed here. Comments are due by 29 July 2024. 

AI and the question of what is "the prior art"?

The international standards for patentability require that an invention is novel and inventive in view of the prior art. The prior art is any disclosure, written or oral, made available to the public before the filing date of the patent (with some grace period exceptions in certain jurisdictions for disclosures made by the inventor). In the pharmaceutical field, for example, the prior disclosure of a specific chemical structure of a drug destroys the novelty of a claim for that drug per se. 

The new power of AI means that it is now possible to generate incapable numbers and variations of chemical structure. These AI generated structures may be included in patent applications or simply published in depositories on the internet. This approach could be used as a strategy by patent trolls or those looking to block others from patenting new structures. As summarised in the USPTO call for evidence: "there is a question whether AI-generated disclosures, especially those with no human input, review, or validation, should qualify as prior art disclosures and potentially preclude human-created inventions from being patented". 

To deal with this problem, a higher enablement standard could be applied to AI generated disclosures, i.e. such that it must be shown that AI disclosures were physically produced, specifically selected and/or otherwise validated with human input. In the USPTO call for evidence, it is suggested that a solution to the proliferation of untested AI disclosures would be to change the rebuttable assumption for this type of prior art such that, unlike other types of prior art, AI disclosures would not be assumed to be enabled. 

Brave new world

Alternatively, AI could be thought of as akin to a natural phenomena, from which isolating a chemical or structure may be considered an invention. The inventive act in this scenario being the isolation of the chemical or structure for a particular purpose. Under existing patent law, the patentability of such an invention would require the patent claim to be purpose limited. Functional claim language of this type is currently more favourably considered in Europe than the US (IPKat). 

Importantly, however, the problem presented by AI-generated prior art for patent law is not qualitatively new. It is just an extension of the existing challenge of non-enabled prior art. This issue will be very familiar to anyone working with method of treatment inventions. In this field, the prior art will often be long lists of possible indications included in prior publications, all of which lack supporting evidence. In Europe, it is already standard practice in this field to argue that such disclosures are not enabled. The proliferation of AI just means that there will now be more of this type of disclosures in other fields. The question becomes whether the rise of AI will therefore encourage and/or permit a shift to more functional language in US patents. 

Another question raised by AI for the definition of prior art, is the definition of disclosure itself. Imagine a scenario whereby an AI is programmed to provide chemical structure and any target suggested by a user. A member of the public is able to ask the AI to provide such a structure, e.g. by inputting the target to an open access website. In this scenario, are the chemical structures "disclosed", whether or not a user has in fact inputted the target? This question is most problematic for UK patent law, where the very potentiality of a disclosure (e.g. an unread thesis in library, or a seed drill tested in a field visible from a deserted footpath) is considered novelty destroying (IPKat).  

On the impact of AI on prior art, the US call of views specifically asks:

1. In what manner, if any, does 35 U.S.C. 102 presume or require that a prior art disclosure be authored and/or published by humans? In what manner, if any, does non-human authorship of a disclosure affect its availability as prior art under 35 U.S.C. 102?

2. What types of AI-generated disclosures, if any, would be pertinent to patentability determinations made by the USPTO? How are such disclosures currently being made available to the public? In what other ways, if any, should such disclosures be made available to the public

3. If a party submits to the Office a printed publication or other evidence that the party knows was AI-generated, should that party notify the USPTO of this fact, and if so, how? What duty, if any, should the party have to determine whether a disclosure was AI-generated?

4. Should an AI-generated disclosure be treated differently than a non-AI-generated disclosure for prior art purposes? For example:

a. Should the treatment of an AI-generated disclosure as prior art depend on the extent of human contribution to the AI-generated disclosure?
b. How should the fact that an AI-generated disclosure could include incorrect information (e.g., hallucinations) affect its consideration as a prior art disclosure?

c. How does the fact that a disclosure is AI-generated impact other prior art considerations, such as operability, enablement, and public accessibility?

 5. At what point, if ever, could the volume of AI-generated prior art be sufficient to create an undue barrier to the patentability of inventions? At what point, if ever, could the volume of AI-generated prior art be sufficient to detract from the public accessibility of prior art (i.e., if a PHOSITA exercising reasonable diligence may not be able to locate relevant disclosures)?

AI and the question of what is "the person skilled in the art"?

The proliferation of AI may also have a significant impact on what constitutes the capabilities of a person skilled in the art. There is no reason why a person skilled in the art would not have access to AI tools available to those in the relevant field. The question becomes therefore, how much these tools permit the skilled person to work an invention or a prior art disclosure. This question has potential implications on the patentability of broad functional patent claims in view of the strict enablement and written description requirements in the US. 

In the pharmaceutical field, for example, the USPTO has a much lower appreciation of the capabilities of a skilled person where it comes to, for example, routine experimentation to find new antibodies binding a known target. In Amgen v Sanofi, for example, the US Supreme Court found that a claim for a genus of antibodies defined by their binding function was not enabled because a skilled person would not be able to work the invention across the whole scope of the claim (IPKat). However, as this Kat has previously opined, it is possible that the enablement threshold of Amgen v Sanofi could be more easily met with the advent of AI technologies for drug design and discovery (IPKat). As it becomes increasingly easier to design and predict the behaviours of new drugs in silico, might we therefore see a lower enablement threshold for functional claiming in the US? In Europe, the EPO already permits broad functional claiming, based on a higher appreciation of the capabilities of the skilled person and a more lenient view of what constitutes "undue experimentation" (IPKat). 

On the impact of AI on the capabilities of the person skilled in the art, the call of views specifically asks:

6. Does the term “person” in the PHOSITA assessment presume or require that the “person” is a natural person, i.e. a human? How, if at all, does the availability of AI as a tool affect the level of skill of a PHOSITA as AI becomes more prevalent? For example, how does the availability of AI affect the analysis of the PHOSITA factors, such as the rapidity with which innovations are made and the sophistication of the technology?

7. How, if at all, should the USPTO determine which AI tools are in common use and whether these tools are presumed to be known and used by a PHOSITA in a particular art?

8. How, if at all, does the availability to a PHOSITA of AI as a tool impact:

a. Whether something is well-known or common knowledge in the art?

b. How a PHOSITA would understand the meaning of claim terms?

9. In view of the availability to a PHOSITA of AI as a tool, how, if at all, is an obviousness determination affected, including when:

a. Determining whether art is analogous to the claimed invention, given AI's ability to search across art fields? Does the “analogous” art standard still make sense in view of AI's capabilities?

b. Determining whether there is a rationale to modify the prior art, including the example rationales suggested by KSR (MPEP 2143, subsection I) (e.g., “obvious to try”) or the scientific principle or legal precedent rationales (MPEP 2144)?

c. Determining whether the modification yields predictable results with a reasonable expectation of success (e.g., how to evaluate the predictability of results in view of the stochasticity (or lack of predictability) of an AI system)?

d. Evaluating objective indicia of obviousness or non-obviousness (e.g., commercial success, long felt but unsolved needs, failure of others, simultaneous invention, unexpected results, copying, etc.)?

10. How, if at all, does the recency of the information used to train an AI model or that ingested by an AI model impact the PHOSITA assessment when that assessment may focus on an earlier point in time (e.g., the effective filing date of the claimed invention for an application examined under the First-Inventor-to-File provisions of the America Invents Act)?

11. How, if at all, does the availability to a PHOSITA of AI as a tool impact the enablement determination under 35 U.S.C. 112(a)? Specifically, how does it impact the consideration of the In re Wands factors (MPEP 2164.01(a)) in ascertaining whether the experimentation required to enable the full scope of the claimed invention is reasonable or undue?

Final thoughts

The question of the potential impact of AI on patentability standards is a far more interesting question than that of so-called AI inventorship (IPKat). AI may have a real impact on what can and cannot be patented, as it becomes easier and cheaper to generate prior art disclosures or to provide examples falling within a broad functional claim (IPKat). In this Kat's view, European patent law is more favourably positioned to tackle this challenge, given its focus on the function of an invention, its lower enablement criteria, and its more permissible novelty standard for selection inventions (IPKat). May the rise of AI therefore see a shift of US patent law towards a more European approach? 

The question raised by the advances in AI demonstrate why patent law is one of the most fascinating areas of law around. More than any other type of law, patent law must evolve in response to technological advancement. The effect AI will have on patentability standards and IP strategy in the coming years, will be fascinating. 

Further reading

• Patenting Antibodies: The epitope claim is dead, long live the epitope claim (Nov 2021)
• Artificial intelligence is not breaking patent law: EPO publishes DABUS decision (J 8/20) (July 2022)
• Sufficiency of broad functional genus and first medical use claims in Europe (T 0424/21) (Oct 2022)
• Bad cases make bad law: Has DABUS "the AI inventor" actually invented anything? (Aug 2023)
• The relevance of G 2/21 to machine learning inventions (T 2803/18) (Aug 2023)
• Broad functional claiming at the EPO (T 0835/21) (Sep 2023)
• Use of large language models in the patent industry: A risk to patent quality? (Oct 2023)