As policymakers and courts continue to grapple with the role of AI in the innovation process, USPTO Director John Squires has issued an Examination Guidance memo regarding AI-assisted inventing. The Guidance rescinds the Office’s prior AI-assisted inventorship framework.

In the memo, the Director acknowledges that the Federal Circuit has held that AI cannot be named as an inventor on a patent application (or issued patent) and that only natural persons can be inventors (citing Thaler v. Vidal, 43 F.4th 1207, 1212 (Fed. Cir. 2022)). Then, he points to traditional inventorship analysis centered on “conception,” commenting: 

“The question is whether the natural person possessed knowledge of all the limitations of the claimed invention such that it is so ‘clearly defined in the inventor’s mind that only ordinary skill would be necessary to reduce the invention to practice, without extensive research or experimentation.’ (citing Burroughs Wellcome Co. v. Barr Labs., Inc., 40F.3d 1223, 1228 (Fed. Cir. 1994)). Analysis of conception turns on the ability of an inventor to describe an invention with particularity.” 

Guidance at 2-3.

Thus, the human who uses AI like any other research tool and who independently understands and can describe the invention with particularity has effectively conceived of it, even if the AI initially assembled the pieces.

You can read the memo in its entirety here: https://public-inspection.federalregister.gov/2025-21457.pdf