The USPTO recently issued Guidance on the patentability of AI. In the US, claims directed solely to a "judicial exception", in the form of an abstract idea, natural phenomena, or law of nature, are not eligible for patent protection (35 U.S.C. 101). However, the case law on subject matter eligibility is infamously difficult to apply. The new Guidance seeks to provide some clarity and includes some practical examples relating to AI and its use in precision medicine.
Legal Background: The existing USPTO guidance on subject matter eligibility
The USPTO has a 2-step approach (Alice/Mayo) for determining whether the subject matter of a claim is eligible for patentability:
Step 1. Is the claim to a process, machine, manufacture or composition of matter? If NO, the claim is not eligible subject matter. If YES, proceed to Step 2A.
Step 2A. Is the claim directed to "a judicial exception", i.e. a law of nature, a natural phenomenon, a product of nature or an abstract idea? If NO: the claim qualifies as eligible subject matter. If YES: proceed to Step 2B. Step 2A is a 2-pronged test. The test of whether the claim is directed to a judicial exception asks whether the claim recites a judicial exception (Prong 1) and whether the claim also does not recite additional elements that integrate the judicial exception into a practical application (Prong 2). If the answer to both is YES, then the claim relates to a judicial exception.
Step 2B. Does the claim recite additional elements that amount to significantly more than the judicial exception? If NO, the claim is not eligible subject matter. If YES, the claim qualifies as eligible subject matter. Readers may be forgiven for finding Step 2A, Prong 1, and Step 2B quite similar. However, Step 2B crucially includes a consideration of whether the additional elements in the claim are "significant", in the sense that they are more than well-understood, routine, or conventional. A claim will be found to be ineligible at Step 2B if the additional elements are found to be insignificant.
Updated Guidance for AI inventions: Step 2A, Prong 1
Prong 1 of Step 2A asks whether the claim in question is directed to "a judicial exception". The Guidance first notes that claims to AI inventions will often involve "abstract ideas" such as mathematical concepts, methods of organising human activity or mental processes. The important question for patent eligibility is whether the claim "recites" the abstract idea, or merely involves or is based on an abstract idea.
![]() |
Before computers |
The Guidance provides some examples of claims that do not recite an abstract idea, including "A treatment method comprising administering rapamycin to a patient identified as having Nephritic Autoimmune Syndrome Type 3 (NAS-3)". In this example, the claim is directed to a practical method of treatment. In implementation, the method requires the mental act of identifying a patient as having NAS-3, which could be achieved using mathematical methods such as an AI-model. However, following the Guidance, these abstract ideas are tangential to the subject-matter of the claim, namely the treatment of a specific disease with a particular compound.
By contrast, a precision medicine approach directed to a method of treatment comprising the steps of a) using an AI model to calculate a risk score to identify patients at risk of a particular disease, and b) treating these patients with an appropriate (non-specific) treatment, is considered to recite abstract ideas (Example 49), in the form of mental acts and mathematical methods.
In another example (Example 47) the following claim is considered: "An application specific integrated circuit (ASIC) for an artificial neural network (ANN), comprising a plurality of neurons organized in an array, wherein each neuron comprises a register, a microprocessor and at least one input; and a plurality of synaptic circuits each synaptic circuit including a memory for storing a synaptic weight, wherein each neuron is connected to at least one other neuron via one of the plurality of synaptic circuits". This claim is also not considered to recite an abstract idea. The claim recites the hardware components formed from a plurality of neurons that function as an ANN. Therefore, while ANNs may be trained using mathematics, the Guidance considers that there is no mathematical concept actually recited in the claim.
Updated Guidance for AI inventions: Step 2A, Prong 2
Prong 2 of Step 2A asks if the claim recites additional elements that integrate the recited judicial exception into a practical application. If the claim as a whole integrates the recited judicial exception into a practical application, the judicial exception no longer applies. For example, the additional elements may improve the functioning of a computer or another technology, link the judicial exception to a particular technological environment or field of use, or use the judicial exception to effect a particular treatment or prophylaxis for a disease or medical condition.
Importantly, it is irrelevant for this test whether the additional elements represent well-understood, routine or conventional activity, this is a question for Step 2B. Instead, according to the Guidance, a claim that "integrates a judicial exception into a practical application" should impose "a meaningful limit on the judicial exception, such that the claim is more than a drafting effort designed to monopolize or pre-empt the judicial exception".
In the precision medicine claim example (Example 49), the claim includes elements in addition to the abstract idea of the AI model. The additional elements include collecting and genotyping patient samples, and administering an appropriate treatment to the patient. However, according to the Guidance, these steps are insignificant extra-solution activities that amount to mere data gathering or, in the case of the treatment step, lack specificity and thus are not meaningful constraints. The additional elements are therefore not considered to recite elements that integrate the abstract idea into a practical application.
By contrast, in the Examples, a claim reciting administration of a specific new compound for the treatment of a specific disease is considered patent eligible. Put simply, administering the new drug is considered an additional element that integrates the judicial exception into a practical application. The practical application arises with respect to the use of the new drug in the specifically identified patient population. In this example, patent eligibility is therefore provided by the answer to Step 2A, Prong 2 and it is therefore not even necessary to consider Step 2B.
Of course, Step 2B provides an additional hurdle to patentability. For Step 2B, it must be shown that the additional elements in the claim are also not well-understood, routine, and conventional in the field. However, Step 2B was not considered in detail in the new Guidance.
Final thoughts
As covered in a separate post, the Guidance on the patentability of AI, is accompanied by separate Guidance on inventorship issues pertaining to AI-assisted inventions (IPKat). As confirmed in both documents, the fact that an invention was developed using AI does not in itself affect the invention's patentability. A key theme from both of the new Guidance documents is the importance of identifying the specific practical applications of an invention involving AI. If they are to be named an inventor, a natural person's contribution to an AI-assisted invention must relate to a particular solution to a specific problem giving rise to the claimed invention (IPKat). Similarly, in inventions relating to the use of AI models, the claim must recite additional elements that somehow integrate the use of the AI model into a specific practical application, such as the administration of a specific compound for the treatment of a particular disease. The USPTO appears keen to avoid broad patents for generalisable uses of AI, applicable in multiple contexts without further inventive input.
Of course, AI precision medicine inventions also face the additional problem that they may be rejected for relating to the judicial exception of a natural phenomena, such as the natural association between a clinical biomarker and the status of disease in a patient. If there is one take away from the Guidance for the field of AI in precision medicine, it is that a lot of care must be taken with both invention capture and claim drafting if the pitfalls of US patent ineligibility are to be avoided.
Further reading
- Artificial intelligence is not breaking patent law: EPO publishes DABUS decision (J 8/20) (July 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)
- Use of large language models in the patent industry: A risk to patent quality? (Oct 2023)
- USPTO call for comments: Impact of AI on patentability (May 2024)
- "Using AI tools to help assess inventive step": A response to the CIPA journal article (June 2024)
- When is the inventor of an AI model also an inventor of the model's output? A closer look at the USPTO Guidance for AI-Assisted Inventions (July 2024)
![New USPTO Guidance on the use of AI in precision medicine](https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEjNIYW1ABmHXZFNm-RmNrOuUeH3LEoq_rF4_QQVNIJl1yET3nMJg9-H2IbNJwPeVVL-rjPtVLBe1P12hPs9qeuqXjcdb0K7sAZEYmFR5aAhvsLbz3nTJ4e6LXCnU3r_l03-sr8jhxwOvCgLAbJ5iwuILvZ58PwPLwxfFNxmHH51i9Yypo-QYAr1/s72-c/Picture1.png)
No comments:
All comments must be moderated by a member of the IPKat team before they appear on the blog. Comments will not be allowed if the contravene the IPKat policy that readers' comments should not be obscene or defamatory; they should not consist of ad hominem attacks on members of the blog team or other comment-posters and they should make a constructive contribution to the discussion of the post on which they purport to comment.
It is also the IPKat policy that comments should not be made completely anonymously, and users should use a consistent name or pseudonym (which should not itself be defamatory or obscene, or that of another real person), either in the "identity" field, or at the beginning of the comment. Current practice is to, however, allow a limited number of comments that contravene this policy, provided that the comment has a high degree of relevance and the comment chain does not become too difficult to follow.
Learn more here: http://ipkitten.blogspot.com/p/want-to-complain.html