The USPTO issued new inventorship guidance for AI assisted inventions on November 28, 2025, relating to the hot topic of AI-assisted inventions. Please note that this only applies to examiners within the USPTO, and not to courts, congress, other countries' patent offices, or WIPO's PCT system.
This patent ain't big enough for a human and AI inventor
While the guidance does not address the harder and more controversial question of can an AI be an inventor or co-inventor of a patent or patent application claim (currently "no" according to the Federal Circuit), it does clarify that a human inventor can use AI as a tool to help them further refine their invention post-conception.
The guidance makes the analogy to laboratory equipment, computer software, or research database being conceptually the same as AI. A "natural person" must have conceived the claimed invention under USPTO AI inventorship guidance. An AI is still not capable of being a patent inventor, consistent with the Federal Circuit's holdings.
The Federal Circuit similarly found that for both design patents and plant patents, they must have a human inventor and that inventor can use AI as a tool.
USPTO patent examiners are instructed to reject patent applications that list an AI system or other non-natural person as an inventor or joint inventor.
Practical steps for complying with USPTO AI inventorship rules
-
- Don't list an AI system as an inventor or co-inventor of your patents and patent applications. Your patent application will be rejected at the very start of examination.
- You can use an AI to further refine your invention after at least one human being conceives of the invention.
- If you can't name any human being that conceived the idea, consider keeping the idea a trade secret (which does not require human involvement).
4. If you can't name any human being that conceived the idea, don't just list a human being as the patent inventor to avoid rejection. This may well arise as an issue in litigation.
Because they contain algorithms, models, processes, and customer insights that provide competitive advantage. When exposed, the loss is immediate and irreversible.
Create an inventory and identify which assets qualify as trade secrets, including code, models, datasets, prompts, and internal tools.
Classification clarifies sensitivity, risk, and required controls. It strengthens governance and ensures teams understand how assets should be handled.
Limiting access by role, tracking interactions, removing outdated permissions, and restricting repositories prevents internal misuse and accidental exposure.
Employees often paste confidential information into public AI tools, causing it to become training data. Governance and private AI tools prevent this.
NDAs, invention assignment agreements, contractor agreements, and offboarding processes ensure confidentiality obligations are clear and enforceable.
Most exposure occurs through human behaviour. Training builds awareness and creates a culture of confidentiality across the organisation.
Audits identify weak points like outdated access, unsecured repositories, shadow AI use, and untracked datasets — key for legal defensibility.
Tangibly provides automated trade secret identification, classification, access tracking, governance workflows, and AI-driven risk detection.