Confluence of equals: AI meets the life sciences

Artificial intelligence (AI) continues to play an increasingly prominent role in life science-related inventions. In the life sciences, AI has been used successfully to predict three-dimensional structures of therapeutic proteins, analyze a patient's genetic predisposition to efficacy across a panel of possible therapeutic treatments, and diagnose medical conditions based on an intricate panel of increasing and/or decreasing biomolecular markers.

These advancements are incredibly valuable; as such, securing meaningful patent protection is critical for the owners of these technologies. Against this backdrop of technological advancement are nuanced issues found in patent law that need to be considered, including how inventorship of AI-derived inventions is treated; how AI influences patent subject matter eligibility; and how AI-generated data can be used effectively in life science-related inventions.

I. Inventorship

A key issue that has been raised based on the role that AI is playing in the invention process is how to address inventorship. The United States Patent and Trademark Office (USPTO) and the United States Court of Appeals for the Federal Circuit have both ruled that inventors listed on a patent application must be natural persons. See, e.g., Thaler v. Vidal, 43 F.4th 1207 (Fed. Cir. 2022). This is consistent with the vast majority of the world's leading patent jurisdictions, which have taken the same view that an inventor cannot be AI per se. For example, the European, United Kingdom, Canadian, and Chinese patent offices have all confirmed that a patent inventor must be a natural person.

II. Patent Subject Matter Eligibility

The inclusion of AI with traditional life science inventive concepts also results in some complexity in terms of what is and what is not considered to be patentable subject matter. Most recently, on July 17, 2024, the USPTO issued a Guidance Update on Patent Subject Matter Eligibility, Including on Artificial Intelligence (Ref) ("Guidance"). The Guidance provides an up-to-date summary of how patent subject matter eligibility should be examined by Patent Examiners. 

Step 1 of the USPTO's subject matter eligibility analysis addresses whether the claimed invention falls into at least one of the four categories recited in 35 U.S.C. 101, namely a process, machine, manufacture, or composition of matter. Step 2 of the USPTO's subject matter eligibility analysis applies the Supreme Court's two-part framework in Alice/Mayo to identify claims that are directed to a judicial exception and to then evaluate if additional elements of the claim provide an inventive concept. Alice Corp. Pty. Ltd. v. CLS Bank Int'l, 573 U.S. 208 (2014) and Mayo Collaborative Servs. v. Prometheus Lab'ys. Inc., 566 U.S. 66 (2012). The Guidance provides a thoughtful analysis of how to carry out the nuanced Step 2 analysis. The Guidance also provides three examples of claims that are helpful to patent practitioners when they are drafting and prosecuting patent applications and wish to have insight into the types of claims that may or may not be patent-eligible.

Of note, Example 49: Fibrosis Treatment provides a helpful analysis into patent subject matter eligibility in the life sciences. Example 49 recites two representative claims: claim 1 focuses on a post-surgical fibrosis treatment method that involves: (a) collecting and genotyping a sample from a glaucoma patient; (b) identifying the patient at high risk of post-surgical inflammation based on a particular AI model; and (c) administering a treatment. Claim 2 depends on claim 1 and introduces the limitation that the appropriate treatment is a defined eye drop.

The analysis does a sound job of: (i) explaining the USPTO's view that claim 1 is ineligible due to step (b) being construed as an abstract idea; and (ii) explaining how claim 2 as a whole integrates the judicial exception into a practical application, and is thus patent eligible.

At the confluence of AI and life sciences, patent subject matter eligibility is very much a nuanced art - but the Guidance provides helpful guideposts for practitioners navigating these murky waters.

III. Data

Another area in which AI has played a significant role in the life sciences space is the rapid generation of meaningful scientific data. For example, in the therapeutic antibody space, the disclosure of amino acid sequence data that demonstrates effective binding between a monoclonal antibody's Fab region and the corresponding epitope on the antigen is critical for securing broad patent coverage. AI has been harnessed to rapidly model how amino acid sequence variants will or will not bind to the epitope; generating this type of amino acid sequence data used to take months to years to complete--now it can be generated in minutes to hours.

An upside to this robust data generation is that it increases the possibility that the patent applicant will be able to secure broad genus claims that cover multiple species of related antibody variants; these types of patents are viewed as being more valuable given that they can effectively prevent design-around options from would-be competitors. In short, the data generation can be used to secure a broader patent monopoly.

However, it is important to note that data generation for the sake of data generation is usually insufficient to warrant broad patent coverage at a patent office; rather, a patent applicant needs to ensure that the generated data correlates with a particular biological or chemical function. For example, linking a particular amino acid sequence--or key residues within that amino acid sequence--with a corresponding change in binding affinity will be far more persuasive to the relevant patent office than just relying on a myriad of potentially interchangeable amino acid sequences as can be the case when AI-generated data is used without further insight or analysis. As with so many aspects of patent law, the devil is in the details. The generation of reams of AI-derived data is really only useful if it is properly tied--through chemical or biological analyses--to the invention in hand.

Conclusion

The confluence of AI and life science inventions continues to expand. As with so many areas of patent law, technological advances are leading the law and the law, at least in the United States, is trying to catch up in an ordered and predictable way.