USPTO asks whether machine-generated disclosures are prior art and whether the skilled person is presumed to use models

2024-04-30USPTO

Request for Comments Regarding the Impact of the Proliferation of Artificial Intelligence on Prior Art, the Knowledge of a Person Having Ordinary Skill in the Art, and Determinations of Patentability Made in View of the Foregoing, 89 Fed. Reg. 34217 (Apr. 30, 2024), Docket No. PTO-P-2023-0044, comments due July 29, 2024

The USPTO posed fifteen questions about how artificial intelligence affects the art against which applications are examined. Question 1 asks in what manner Section 102 of Title 35 of the United States Code presumes or requires that a prior art disclosure be authored or published by humans, and how non-human authorship affects its availability as prior art. Question 4 asks whether a disclosure generated by artificial intelligence should be treated differently for prior art purposes, including by degree of human contribution, hallucination, operability, enablement and public accessibility; the notice observes that a public disclosure is presumed enabling and asks whether disclosures "that have not been prepared and reviewed by a human" should receive the same rebuttable presumption. Question 5 asks when the volume of generated prior art could become an undue barrier to patentability, and the notice records commenters' view that artificial intelligence could generate a vast amount of prior art. Questions 6 through 10 ask whether the person having ordinary skill in the art must be a natural person, which tools that person is presumed to know, and how their availability affects common knowledge, claim interpretation, analogous art, reasons to modify, predictability and objective indicia. Question 15 asks whether Title 35 should be amended. The notice states that it "is not a guidance document."

What it changesNothing changes in examination today, but the questions show where the obviousness bar could move for incremental artificial intelligence work. If the skilled person is presumed to use the tools in common use, an improvement a model could suggest on request, such as a new loss weighting, a standard architecture applied to a new modality or a routine hyperparameter search, becomes harder to defend as non-obvious, and the argument turns on what the tools of the filing date could and could not do. The response for a filing program is to put the non-obvious human judgment on the page: the problem the obvious approach did not solve, the result it would have predicted, and the unexpected result actually obtained, with the data. Programs that rely on defensive publications should note the open question whether machine-generated text is presumed enabling, and should keep publishing human-reviewed disclosures with working detail. Comments are due July 29, 2024.

More on Prior art, public disclosure and obviousness

← All recent developments · Search all developments

Burbage Law is the practice of Jon Burbage, a registered patent attorney. 342 United States patents have issued from applications he wrote; the full list is on the Patents page. More about Jon Burbage.

This is general information about a published decision, agency guidance or piece of legislation. It is not legal advice and does not create an attorney-client relationship. Items concerning offices outside the United States are reported for information only and are not counsel on the law of those countries.