AI cannot be listed as inventor, human element a must, USPTO says

The US Patent and Trademark Office (USPTO) has issued its official stance on protections for AI-assisted inventions – and human contribution is the key.

The USPTO released a preview of its new inventorship guidance on Monday to help provide clarity to agency stakeholders and personnel as “AI systems play a greater role in the innovation process.”

The new guidance will help define the parameters for those who seek to file patent or register trademark protections for inventions assisted by artificial intelligence, including those assisted by generative AI.

“This guidance explains that while AI-assisted inventions are not categorically unpatentable, the inventorship analysis should focus on human contributions, as patents function to incentivize and reward human ingenuity,” the USPTO stated in its 27-page report on the matter.

Specifically, the agency says that patents will only be granted for inventions in which a “natural person” has provided a significant contribution.

The USPTO, who wants feedback from the public, has also come up with a set of procedures to help determine the parameters, such as defining the different ways a human being can contribute to the conception of an AI-assisted invention.

Entities beholden to the new edict will include the USPTO’s Central Reexamination Unit and the Patent Trial and Appeal Board.

Finally, the new guidance, which becomes effective February 13th, will discuss the impact the procedures have on other aspects of patent practice.

“Promoting responsible innovation, competition, and collaboration will allow the United States to lead in AI and unlock the technology’s potential to solve some of society’s most difficult challenges,” the USPTO said.

In development since 2019

The USPTO has been working on developing a guideline for AI-assisted inventions since 2019, issuing numerous requests for comments, holding panels, meetings, public listening sessions, and publishing several reports of its findings.

One requirement put forth by the USPTO is that each human inventor must take an oath or declaration to claim inventorship.

An application that names an AI system as the sole inventor will not be accepted, the USPTO states.

Another issue taken up by the USPTO is the use of AI-cited research from erroneous sources – considered a serious act of misconduct when presenting a case to the patent or trademark appeal board.

Quoting US Supreme Court Justice John Roberts, USPTO Director Katherine K. Vidal wrote in a February 6th Memorandum, “AI has well-recognized shortcomings, including being prone to present inaccurate or nonsensical information as fact, a phenomenon referred to as ‘hallucinations.’”

“Risks are heightened when new technologies are employed without full appreciation of the shortcomings of those technologies. Simply assuming the accuracy of an Al tool is not a reasonable inquiry,” Vidal stated.

To address this issue, Vidal stated “any paper submitted to the USPTO under signature must be reviewed by the person presenting the paper. ”

Vidal, also the Under Secretary of Commerce for Intellectual Property, said any submission – including an AI-generated or Al-assisted – misstating facts or law could be construed as "a paper presented for an improper purpose because it could ‘cause unnecessary delay or needless increase in the cost of any proceeding before the Office.’”

The person presenting a paper based on "hallucinations" could be subject to a variety of disciplinary sanctions and even criminal liability, the memo said.

Public comments must be submitted within 120 days and through the Federal eRulemaking Portal, entering docket number PTO-P2023-004.

The USPTO national registration authority is a division of the Department of Commerce.