Patent Office Releases Role of AI in Inventorship Determinations

In February, the PTO released “Inventorship Guidance for AI-assisted Inventions (89 CFR 10043) and noted that the Office will issue examples to assist in the examination of applications in accord with this Guidance. I recently finished reading and annotating the Guidance and the Examples, that amounts to a short course on inventorship in all its fact-driven glory. The Guidance goes beyond factors involving the determination of inventorship, and tackles the duties of disclosed in Rule 56 and the duty of reasonable inquiry and the requirements for information. The latter can be made by the examiner “or other USPTO employee” (37 CFR s. 1.105). This can include an inquiry into inventorship and the role of AI in making the invention. Remember, the application cannot name an AI inventor, since AI is not a natural person, but the role that AI played in making the claimed invention may well be relevant to the role(s) played by the named inventors, and that role may not meet the “significant contribution” required by the Pannu standard.(Pannu v. Iolab Corp., 155 F.3d 1344 (Fed. Cir.  1998)). Put another way, “[A] new inventor cannot be named if no natural person made a significant contribution to an AI assisted invention” (89 C.F.R. 10048). So AI can “invent”, just not be named as an inventor on any claim when the application is filed.

Since the Examples lean heavily on the Pannu factors, here they are: (1) Each inventor must contribute in some significant manner to the conception or reduction to practice of the invention; (2) make a contribution to the claimed invention that is not insignificant in quality when that contribution is measured against the dimension of the full invention; and (3) do more than merely explain to the real inventors well-known concepts and/or the current state of the art.

The attempt to define “significant” and to explain the duo of “conception” and “reduction to practice” take up about six columns in the CFR guidance, and are defined mostly in terms of negative activities. However, since the examples that I will try to briefly outline result in a lead compound, there is the one mention of a bioactive drug:

“’In some instances an inventor may only be able to establish a conception by pointing to reduction to practice by a successful experiment.’ This concept of simultaneous conception and reduction to practice is sometimes pertinent in unpredictable arts, where, for example, the inventor does not have a reasonable expectation that they would produce the claimed invention. Under those circumstances, the conception of a specific chemical compound does not occur until the reduction to practice … occurs. [Pannu factors do] not imply that reduction to practice is sufficient for invention or a substitute for conception.”

This leads in to a discussion of the three hypothetical claims in Example 2. In the first, Professor Marisa (“M”) wants to ascertain if compounds that block the mutated androgen receptor might be effective to treat prostate cancer and exhibit few side effects. She enlists the aid of Raghu (“R”), and asks him to generate compounds that will block the receptor. He uses a pre-existing AI system that measures binding of a target, the androgen receptor, to a large group of compounds (20,000) provided by the NCI. From the output, M chose to do further work with six compounds that exhibited the best binding. One compound, CD_1, exhibited the best binding.

Now the post doc, Naz, enters the picture. I have created a small table to summarize M and N’s contributions in Scenario 1, since the text of the example is unclear in some places:

N	Joint Effort	M
Synthesized M’s top six		Identified the top six
Identified CD-1 as the best binder
	Change the structure of CD-1 so binding is more selective
		Select the “preferred alterations in CD-1
Naz makes stable analog
		M makes selective potent analog of Naz’ analog
	M’s analog is a lead compound (Still joint? PTO says “Yes.”)
		M appreciates utility of AI/lab method (completes conception)

Claim 1 in application as filed is a general method claim that does not claim M’s compound but recites AI screening and Naz’s contribution of making a stable analog. PTO says that this is sufficient for them to be co-inventors. Raghu is not a co-inventor.

Claim 2 is a compound claim. Likewise, the PTO finds that M and Naz are co-inventors—each contributed significantly to the claimed compound.

In Claim 3, Raghu is posited as having to do more on the AI system than simply feeding it 20,000 compounds, but is involved in “designing, building or training” the system, so Raghu joins M as a co-inventor.

  Back to All Resources