USPTO strategy confirms examiners search by machine similarity and promises guidance beyond eligibility

2025-01-14USPTO

United States Patent and Trademark Office, "Artificial Intelligence Strategy" (Jan. 14, 2025), announced in "USPTO announces new Artificial Intelligence Strategy to empower responsible implementation of innovation," uspto.gov (Jan. 14, 2025)

The United States Patent and Trademark Office published an Artificial Intelligence Strategy organized around five focus areas: developing intellectual property policies that promote inclusive innovation and creativity in artificial intelligence; building computing infrastructure, data resources and product development for the Office's own artificial intelligence tools; promoting responsible use of artificial intelligence inside the Office and across the innovation ecosystem; building artificial intelligence expertise in the workforce; and collaborating with other agencies, international partners and the public. The strategy reports that nearly 80 percent of patent examiners had used artificial-intelligence-powered search features such as More-Like-This-Document and Similarity Search, and it describes tools already in use that help classify nonprovisional utility applications and retrieve potential prior art for examiners. On policy, it states that the Office is developing guidance on how artificial intelligence bears on subject matter eligibility, obviousness, enablement, written description and inventorship, building on the February 2024 inventorship guidance and the July 2024 eligibility update. On the workforce, it commits to recruiting examiners with artificial intelligence backgrounds and to training examiners and Patent Trial and Appeal Board judges on the technology and on responsible use of the tools. The strategy is a planning document; it changes no rule, form or examination procedure.

What it changesMost examiners now run machine similarity searches over the text of the application, not only keyword queries built from the claims, so the nearest reference is increasingly found by what the specification says. Run the same kind of search on the draft before filing, read the top results, and write the background and the detailed description to distinguish them in the mechanism the claims recite; a specification written in the vocabulary of its closest prior art will be matched to that art. The promised guidance on obviousness, enablement and written description is the next front after eligibility for artificial intelligence applications, and a specification cannot be supplemented after filing. Applications filed now should already carry what that guidance will test: the model architecture, how it was trained and on what kind of data, and at least one worked example with measured results. The strategy itself binds no examiner.

More on Prior art, public disclosure and obviousness · Eligibility of software and AI claims

← 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.