On April 30, 2024, the USPTO launched an RFC (Request for Comments) inviting the public to comment by July 29, 2024 on the issue to know whether content generated by generative tools may be considered as prior art, and the impact such tools may have on the notion of a person having ordinary skilled in the art (PHOSITA).

This RFP is the second step in a process begun by the USPTO on February 13, 2024, when it adopted Inventorship Guidance for AI-Assisted Inventions, in which the USPTO recognized that such inventions may give rise to the grant of a patent, provided that the human being has made a significant contribution and that the grant of the patent appears to be a justified reward for such effort. With the rapid progress of these systems, the question of where the cursor should be placed on this human contribution remains open.

While the use of an artificial intelligence system as an assisting tool does not, in my view, rightly exclude the granting of a patent, the question remains as to whether the state of the art should include content generated using generative tools. This is the purpose of the RFP launched on April 30, 2024.

Far from being an academic issue, the question is all the more acute given that some companies, such as All Prior Art, make a point of generating enforceable disclosures through such tools. In his blog, Ryan N. Phelan highlights the central role played by human beings in the patent law system, and the importance attached by US courts to the fact that invention includes the idea of “conception”, which requires a mental step specific to human beings (Univ. of Utah v. Max-Planck-Gesellschaft Zur Forderung Der Wissenschaften E.V., 734 F.3d 1315, 1323 (Fed. Cir. 2013)). From this, Ryan N. Phelan deduces that such prior art might not be enforceable for lack of having been “conceived” by a human being.

However, this overlooks the fact that this case law was issued in relation to the question of whether an invention is patentable, which patentability requires a human contribution. However, a disclosure may be part of prior art without being patentable, and therefore without necessarily testifying to a human “conception”.

Given that the novelty of the invention has to be assessed objectively on a worldwide scale, we may well ask whether such content generated by generative tools should not “objectively” form part of the state of the art. At most, the question could arise as to whether prior art which is not sufficiently “reduced to practice”, to use the American terminology, and which does not enable a PHOSITO to execute it, should be enforceable.

The development of these tools therefore raises fundamental questions about the innovation process and the role played by patent law in this respect. At a time when innovations may increasingly be the result of the use of generative tools, the fundamental question arises as to what should or should not be protected, and under what conditions. Should we rethink the concept of novelty? Or that of the PHOSITO? Should patents only be granted to those who not only “conceive” the invention, but are also in a position to exploit it industrially? The questions are manyfold, and so are the possible answers. These answers and their implementation will have a considerable impact on innovation processes. It’s up to each and every one of us to contribute to this effort of reflection, which certainly deserves a global and collective approach.