USPTO's Revised Inventorship Guidance Explained
The USPTO’s revised guidance does not treat AI-assisted inventions as categorically unpatentable, but it requires each claim to have at least one human inventor who made a significant contribution. Applicants must disclose inventorship accurately, identify the humans who conceived the claimed invention, and avoid naming an AI system as an inventor. Under Thaler v. Vidal, an AI cannot be an inventor, so disclosure practices should document how humans shaped the conception and reduction to practice rather than merely prompting or supervising an AI tool.
Also worth reading: Can AI-Assisted Inventorship Errors Still Be Corrected Under Section 256? · What Should an AI-Assisted Inventorship Documentation Checklist Include? · AI Patent Inventorship Requirements After USPTO Guidance?
These rules affect prosecution strategy and risk management. Inventors and companies should keep contemporaneous records of AI use, human contributions, and ownership; careless or incomplete inventorship disclosures can jeopardize validity or trigger inequitable conduct concerns. Disclosure to generative-AI tools can also create confidentiality and privilege risks, so counsel should limit sensitive submissions and review outputs. Overall, AI-assisted inventions remain patentable, but only with accurate human-inventorship disclosure and a defensible record.
Thaler v. Vidal and Human Inventors
The Federal Circuit’s decision in Thaler v. Vidal confirms that only natural persons may be named as inventors on a U.S. patent, meaning AI systems cannot legally hold inventorship. For AI-assisted inventions, this forces applicants to identify the human(s) who conceived the claimed invention with the aid of AI tools. Disclosure rules therefore matter: applicants must accurately state each human inventor’s contribution while avoiding false claims about AI’s role. Failure to do so can jeopardize patent validity or expose prosecution risks.
Recent USPTO guidance and global developments, including China’s Supreme Court AI rules and WIPO’s SME resources, emphasize transparency and human oversight. Practically, inventors should document how AI was used, who directed it, and how the final conception occurred. Because disclosing AI use to generative tools may create prosecution risks, counsel often advise careful records and privilege protections. Ultimately, AI-assisted inventions remain patentable only when a human inventor is properly identified and the disclosure rules are satisfied.
Disclosure Risks with Generative AI Tools
How Do AI Patent Inventorship Disclosure Rules Affect AI-Assisted Inventions? Under current USPTO guidance, only natural persons may be named as inventors, so a human must have contributed significantly to every claim. Yet disclosing that generative AI helped draft, refine, or conceive aspects of an invention can trigger prosecution risks, because examiners may question whether the human contribution meets the bar for conception. Applicants who overstate AI’s role risk invalidating claims; those who conceal it risk inequitable conduct if the tool’s use later surfaces.
The safer path is documenting human inventive contribution before using AI, limiting prompts to routine drafting, and disclosing only what the rules require. Thaler v. Vidal confirms that AI cannot be an inventor, while recent China Supreme Court rules and WIPO guidance signal global convergence on human-centered inventorship. Patent lawyers increasingly act as shields, vetting AI-assisted disclosures to protect inventions from misuse and preserve enforceability.
Global Perspectives on AI Inventorship
Disclosure rules for AI-assisted inventions have become a critical pressure point in patent practice worldwide. In the United States, the USPTO's revised guidance confirms that AI systems cannot be named as inventors, but it also requires applicants to disclose the significant contribution of human inventors when AI tools play a substantial role in creation. This means practitioners must carefully document how human ingenuity shaped the claimed invention, since failure to identify natural persons who made meaningful intellectual contributions can render a patent invalid or unenforceable. The Thaler v. Vidal decision reinforced this framework by holding that inventorship requires a natural person, leaving AI-generated outputs without protection unless a human can demonstrate significant contribution.
Elsewhere, the landscape diverges considerably. China's Supreme Court has issued judicial rules addressing AI-related disputes, signaling a more accommodative stance toward technology-driven innovation while still anchoring inventorship in human actors. Meanwhile, disclosure of confidential invention details to generative-AI tools creates its own prosecution risk, potentially triggering loss of novelty or trade secret protections. For small and medium-sized enterprises navigating this terrain, as WIPO's educational materials emphasize, understanding both inventorship disclosure obligations and the risks of AI tool usage is essential to preserving patent rights across jurisdictions.
Practical Steps for AI Patent Applicants
Under current USPTO guidance, only natural persons can be named as inventors, so an AI system cannot be listed even if it contributed to the conception of an invention. Applicants must therefore disclose how AI was used during the inventive process and identify the human inventor(s) who conceived the claimed subject matter. Failure to do so can jeopardize patent validity, as seen in Thaler v. Vidal, where the Federal Circuit confirmed that inventors must be humans.
Disclosure also carries prosecution risk. Statements about generative-AI tools can be used against applicants if they suggest the human contribution was insufficient or if confidential inputs were exposed. China’s Supreme Court and WIPO have issued related guidance emphasizing transparency and human-centered inventorship. Practically, applicants should document AI-assisted workflows, avoid over-disclosing proprietary prompts, and ensure at least one human meets the conception standard before filing.
AI Inventorship Rules: US vs. Global Approaches
| Approach | Disclosure & Inventorship Rule | Effect on AI-Assisted Inventions |
|---|---|---|
| United States (USPTO; Thaler v. Vidal) | AI cannot be a named inventor; a natural person must make a significant inventive contribution. AI use is not categorically disclosed, but duty-of-candor and inventorship accuracy matter. | AI-assisted inventions can be patented with sufficient human contribution; incorrect inventorship or undisclosed AI-related facts can create prosecution and validity risk. |
| European Patent Office | Inventor must be a natural person; AI cannot be inventor. AI use is not generally separately disclosed unless relevant to inventorship or patentability. | Applicants should document human inventive contribution and ensure inventorship is accurate across AI-assisted filings. |
| China (CNIPA/Supreme Court guidance) | Human inventorship is required; AI contribution and disclosure may become evidence in disputes over inventorship, ownership, or validity. | Strong documentation of human contribution is important; vague AI-use records can weaken enforcement or increase correction risk. |
| WIPO/SME global guidance | No harmonized global disclosure rule; most systems require human inventors and vary on AI-use disclosure and evidence. | Cross-border AI-assisted filings need jurisdiction-specific disclosure strategies and consistent records of human inventive steps. |