Only real people can patent inventions — not AI — US government says
![<i>skynesher/E+/Getty Images</i><br/>Billions of dollars in investments will guide how artificial intelligence is marketed and used.](https://kion546.b-cdn.net/2024/02/cnn-L19jb21wb25lbnRzL2ltYWdlL2luc3RhbmNlcy9jbHNrc21pYnMwMDA1M2I1d3ZzZ21rY2Rq-L19jb21wb25lbnRzL2FydGljbGUvaW5zdGFuY2VzL2Nsc2tveDkybTAwMjhrbnBjaDF6bWV6MTg.jpg)
Billions of dollars in investments will guide how artificial intelligence is marketed and used.
By Brian Fung, CNN
(CNN) — Artificial intelligence promises to supercharge scientific research and entrepreneurial innovation, helping researchers and inventors make new discoveries and creations.
But how will patent protections apply to inventions made with the help of AI tools, particularly generative AI?
For the first time, the US government has provided an answer, one that will shape how everyone from large businesses to home tinkerers alike can apply for intellectual property protections.
The decision could influence the future of billions of dollars in investments and subtly guide how artificial intelligence is marketed and used.
On Tuesday, the US Patent and Trademark Office (USPTO) said that to obtain a patent, a real person must have made a “significant contribution” to the invention and that only a human being can be named as an inventor on a patent.
The official guidance published this week provides a boost to innovators by reassuring them that their inventions involving AI can be patented, while continuing to enshrine human creativity and ingenuity by establishing basic expectations about how AI could make or break a patent application.
The guidelines reflect the Biden administration’s swift moves to get ahead of artificial intelligence issues. In a sweeping executive order last fall, the White House had directed the USPTO to issue its inventorship guidelines by the end of February.
Precisely what constitutes a “significant contribution” is a little bit fuzzy and case-specific, and much of how the guidelines work will necessarily be figured out in real-time as they are applied and, perhaps in some situations, debated in court.
“The challenge will be in implementing the guidance,” said Jamie Nafziger, an attorney at the law firm Dorsey & Whitney. “How sophisticated of a prompt will be required for a given invention? In connection with training an AI system, what level of planning will be required? Patent examiners will surely have some interesting challenges ahead.”
Still, the overall picture the USPTO has painted is one where real people continue to sit at the center of the US patent system, in what experts describe as a logical extension of the status quo.
How inventors can still patent AI-assisted creations
The USPTO has provided some hypothetical examples of ways that its guidelines could work. For instance, an inventor who simply asks an AI chatbot to design a critical part for a remote-control car would not be eligible for a patent on the car, because he or she didn’t make enough of a contribution to the car’s invention. It was the AI that did the hard part, not the human.
“A natural person who only presents a problem to an AI system may not be a proper inventor” of something, the USPTO said. “However, a significant contribution could be shown by the way the person constructs the prompt in view of a specific problem to elicit a particular solution from the AI system.”
If an inventor could show the work he or she did to get the AI chatbot to produce a specific design that enabled the remote-control car to function, then that could open the door to a patent, according to the USPTO.
The USPTO guidance builds on existing case law. A federal appeals court already held last year, in a case known as Thaler v. Vidal, that only actual people can be listed as inventors on US patents, effectively ruling out the possibility for AI to be named as an inventor or co-inventor.
In that case, the USPTO had rejected the patent applications of an inventor who had given his AI system sole credit for the inventions.
The inventorship guidelines will help everyone understand the boundaries of patent protections as AI is increasingly used in the inventive process. That may resolve some of the uncertainties that could otherwise slow the development and use of AI, said Randy McCarthy, an attorney at the law firm Hall Estill.
It’s also consistent with how the US Copyright Office is approaching copyright protections in relation to AI, McCarthy added.
“Some sort of human agency is required, or no protection is available,” he said. “A practical result is that, when creating a new design, artwork, invention, novel, or even computer code, a human can use an AI-based system to assist in the creation of this content, but must make sure that they are sufficiently involved in the process.”
At the same time, the USPTO guidelines don’t require inventors to disclose the use of AI, and some worry it could encourage so-called patent trolls to apply for broad patents that don’t lead to any actual creations but serve as the foundation for bogus patent lawsuits.
“The economy already is harmed by a surplus of low-quality patents which leads to unproductive litigation, rent-seeking and transfers of wealth from productive businesses to those who are experts at navigating the legal system,” said John Bergmayer, legal director at the consumer advocacy group Public Knowledge. “My worry would be that AI-assisted “inventions” [where the involvement of AI might be concealed] or just AI-assisted patent applications supercharges this.”
The-CNN-Wire
™ & © 2024 Cable News Network, Inc., a Warner Bros. Discovery Company. All rights reserved.