The USPTO issues new guidelines clarifying that generative AI is considered a tool that assists human inventors and cannot itself be considered an inventor
What to do about AI inventorship is a hard problem. … Bluesky: Mark Lemley / @marklemley : What to do about AI inventorship is a hard problem. The prior PTO guidelines, now withdrawn, at least tried to grapple with it. The current approach is simply to say that people can be inventors, AI can't, and assume that ignoring AI conception somehow takes care of the problem. It doesn't. … Mark Lemley / @marklemley : In practice, I suspect this means applicants will lie about who made AI-generated inventions, the PTO will let them, and those patents will be in trouble if and when they are enforced in court. — 3/3 Forums: Msmash / Slashdot : US Patent Office Issues New Guidelines For AI-Assisted Inventions
Context & Ripple Effects
The guidance formalizes a position already reflected in a federal ruling that patent inventors must be human and earlier European rejections of applications naming AI as inventor. It shifts the unresolved issue from whether AI can be listed to how human conception is established when generative tools contribute to an invention.
That framing parallels the Copyright Office's earlier focus on whether AI-assisted work reflects a person's own mental conception. The practical significance is therefore evidentiary: applicants retain access to patents, but must connect the claimed invention to a human inventor.
First-order effects
- Patent applicants using generative AI must name human inventors; AI is treated as an assisting tool rather than a rights-bearing inventor.
- The USPTO's approach preserves the existing human-inventor filing framework while leaving applicants to document the human role in conception.
Second-order effects
- Patent counsel and R&D teams will face greater pressure to maintain records of prompts, review, experimentation, and human decision-making where AI materially contributed.
- Challenges to AI-assisted patents may increasingly focus on inventorship accuracy and proof of human conception, rather than on an attempt to name the model itself.
Third-order effects
- If AI becomes more central to technical discovery, patentability may hinge less on the tool's use than on a workable standard for separating human conception from machine-generated output.
- The policy direction keeps patent law anchored to human authorship, but persistent disputes over disclosure and attribution could force more explicit rules or case-law tests.
The trend: AI governance is moving from categorical debates over whether systems can hold legal status toward operational rules for proving accountable human involvement.