Artificial Intelligence (AI) is a technology that promises to reshape our everyday lives. How we approach work, school, and play will change considerably as AI technology continues to integrate into our society. Earlier this year, the United States Patent and Trademark Office (USPTO) requested public comment on AI and inventorship to determine if statutory or regulatory changes are necessary to address AI contributions to inventions, as well as if there is any legal or policy precedent that could guide the USPTO’s treatment of AI inventors. Request for Comments Regarding Artificial Intelligence and Inventorship, 88 Fed. Reg. 9492 (Feb. 14, 2023). The purpose of the request is to “foster and promote AI-enabled innovation.” However, there is doubt whether the current statutory framework has this ability.

Two immediate issues for AI-conceived inventions are inventorship and assignment of patent rights. According to the Federal Circuit, “the Patent Act requires an ‘inventor’ to be a natural person.” Thaler v. Vidal, 43 F.4th 1207, 1209 (Fed. Cir. 2022). Even if AI could be listed as an inventor, there is currently no legal mechanism by which an AI machine could assign the rights to the technology it conceives.

One potential workaround to the first issue is to list the natural person who created inputs or caused the input of data into the AI machine, or otherwise assisted the AI machine in conception of the invention, as the inventor. In Thaler, the court explained that it was not confronted with “the question of whether inventions made by human beings with the assistance of AI are eligible for patent protection,” and so did not resolve that issue. Id at 1213. Entities will have to hold off on employing this workaround until administrative and judicial interpretations of the Patent Act catch up to the capabilities of generative AI.

It is also unclear what ramifications AI-conceived inventions will have on the patent system once they are eligible for patent protection. In theory, AI machines will have inventive capabilities more advanced than those of their human counterparts. AI dominance in the field of inventorship could marginalize human inventorship and give companies who own the best generative AI an unfair advantage in their ability to construct patent walls and box out competition. Further, AI could conceivably become so adept at producing patentable inventions that the USPTO would become oversaturated with applications, leading to extended prosecution periods.

As the uncertainties surrounding AI-generated inventions increase, it may become necessary for Congress to proactively amend the Patent Act. While there is no silver bullet, and adequately molding the patent system to accommodate generative AI will likely be an iterative process, there are some legal concepts that could help smooth the transition.

One legal mechanism that could be helpful, and is already available in a number of jurisdictions, is the utility model. At least seventy-five countries recognize and grant some patent protection to incremental inventions known as “utility models.” Utility models are inventions that meet the “novelty” requirement for patent protection but fail the inquiry for “obviousness” or “inventive step.” In exchange for less stringent patentability requirements, utility models offer a shorter duration of patent protection. Typically, there is little to no substantive review of applications in a utility model system; substantive review is left to administrative agencies or the courts in the event of litigation.

The United States can emulate the utility model example by creating a patent system where human-made inventions would continue to undergo substantive examination and have lengthy protection periods, while AI-conceived inventions would be procedurally granted and carry shorter periods of protection. This sort of system would alleviate two of the biggest concerns relating to AI-conceived inventions. First, it has the ability to provide greater protection to human-conceived inventions and maintain their importance in the marketplace. Second, it can prevent a backlog in examination time at the USPTO by bypassing substantive review for AI-conceived inventions.

Utility model application in the United States could also be constructed to mirror the framework of several European countries by affording patent protection to human-conceived inventions that are deemed novel but obvious. This would make patent protection more accessible to smaller inventors who cannot afford the current substantive patent review process in the United States. Larger entities would also benefit, as they would have more options when seeking to protect their innovations. Utility models are also more adept at protecting inventions in fields where the innovation rate is high, as protection can be obtained quickly at a lower cost. This would help ensure patent protection continues to be incentivized as the technological landscape continues to evolve.

In an ideal world, the Patent Act would be flexible enough to fold AI inventorship into the patent system seamlessly. However, the reality is that the novelty and ability of AI will likely continue to create ambiguity and uncertainty for entities seeking patent protection as the legal framework continues to evolve.

This publication is intended for general information purposes only and does not and is not intended to constitute legal advice. The reader must consult with legal counsel to determine how laws or decisions discussed herein apply to the reader's specific circumstances.