Contesting American Patent Law with AI Inventions
While many fields, such as healthcare, financial services, and manufacturing, are seeing increased efficiency and discoveries due to the implementation of AI, some areas are contested about what the existence of AI means moving forward. The influence of AI in the inventive sphere has called the traditional patent law practices into question, instigating the debate of whether AI can be applied to customary patent law or if a new, refined law has to be established to accommodate AI-inventors.
In 1952, the modern day patent law was enacted in order to refine the prior loose patent system. The “invention” requirement was the staple of traditional patent law. This stated that the proposed device had to solely appear to be ingenious or beyond the typical idea that one might devise. Due to lack of usefulness or novelty, many courts tended to invalidate patents and a resolution to the original law was sought out as a solution. [1] In 1952, a new law of patent jurisdiction was created. Title 35 of the U.S. Code of the Patent Act addresses patent law by stating that a patent can be secured by “Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent thereof, subject to the conditions and requirements of this title.” [2] The Patent Act was created under the impression that inventors would always be human and in turn, only humans qualify as inventors. As there was a need for reform in 1952, there is now a new resurgence for reform as humans began to relinquish their inventorship to the robotic hands of AI inventors.
In the Yale Journal of Law and Technology, “Artificial Intelligence and Innovation: The end of Patent Law as We Know It,” Tim W. Dornis explains the extent of AI capabilities in the modern era: “AI is able to ‘invent’ not only new materials and machines but also manufacturing processes, pharmaceutical drugs, and household products… [humans] are no longer the masters of innovation.” [3] The responsibility of inventing that was previously held by humans is being passed on to AI. In order to adapt to this new technological approach in the world of innovation, a United States patent doctrine was created which states that the first human to “recognize” an invention that is generated by AI becomes the “inventor” when dictating origins in patent law. [4] This is the first step to determining how to classify AI inventions since this is a rapidly evolving field. While this clarifies who receives credit for inventions, there is still an implication that only humans can be inventors. In addition, following solely this doctrine leads to potential problems with ownership, especially when the human who “recognizes” the invention is not the creator of the AI. In November of 2025, the United States Patent and Trademark Office released a statement explaining that an inventor must be human and AI is simply an instrument that cannot “qualify for inventor status.” [5] While this definition withstands current AI practices in inventorship, it is only temporary due to the unknown nature of future AI machines. There is a need for preparation of future legislative classification of patents because of the trajectory of AI capabilities.
While AI is at present unable to autonomously establish direct research questions and invention goals, it is able to provide the conception of an invention. This is the base of the current conflict around AI inventions. AI has a sense of autonomy in its creations since “conception [is viewed] as the touchstone of inventorship.” [6] AI’s autonomy leads to a lack of human involvement, creating the paradox of “an invention without an inventor.” [7] This poses a problem for the customary approach to patent law since an inventor is presumed to be a human. AI itself is a paradox since it is an invention that has become an inventor-adjacent.
There are two ways to solve this complication. The first would be to eliminate patents altogether since the extensive implementation of AI would make keeping it out of the inventive sphere next to impossible. The alternate choice to the elimination of the patent law is adaptation to a new method.
Patents revolve around solidifying inventions that were created to progress and assist humans in every area from day-to-day life to advanced technological problems. However, Emna Chikhaoui and Saghir Mehar, law professors at Prince Sultan University in Saudi Arabia, note that there is a disconnect between the inventive motives of humans and AI autonomy in their article, “Artificial Intelligence (AI) Collides with Patent Law.” Chikhaoui and Mehar explain that, “[this is] a situation where the behaviour of computers for the purposes of artificial intelligence is autonomous and is not supported by human decision-making.” [8] Chikhaoui and Mehar propose a model that could solve the current conflict between AI and the patent process, in addition to future conflicts that could arise. In order for a model to be successful, it needs to incorporate the human ability to override biases that may be implemented by the AI. In their model they address the three components of patents: novelty, utility, and non-obvious. [9] For Chikhaoui and Mehar, Novelty is handled by AI, since it can rapidly search data bases to ensure novelty exists. Utility is under human control since inventions and processes are made to provide for humans. Testing the non-obvious nature of a product is designated to both humans and AI, due to the nature of this aspect being more difficult to objectively determine. After these three steps, there is a final check by humans before the invention can be submitted as a patent. This process incorporates AI and human steps in order to provide a specific patent owner as well as using AI in a productive manner.
With the continual development and administration of AI it is important to understand the implications in these changes of lifestyle. Patents are widely respected and recognized in the United States and as AI-inventors affect how ownership, validity, and patent protection are defined and enforced. AI inventions are already being applied in engineering and medical fields. The question of if AI-inventors deserve the same protections as human inventors will be a very important one as AI inventions continue to be integrated into society.
Sources
Taylor, David O. "Patent Reform, Then and Now." Michigan State Law Review 2019, no. 2 (2019): 431-510. HeinOnline.
35 U.S. Code Pub. L. No. 110, 66 Stat. 797 (1952).
Dornis, Tim W. "Artificial Intelligence and Innovation: The End of Patent Law as We Know It." Yale Journal of Law and Technology 23, no. 1 (2020): 97-159. HeinOnline.
Ibid.
United States Patent and Trademark Office. “Revised inventorship guidance for AI-assisted inventions.” Last modified November 26, 2025 at 09:11 (EST).
Dornis, “Artificial Intelligence and Innovation,” 2020.
Ibid.
Chikhaoui, Emna, and Mehar, Saghir. "Artificial Intelligence (AI) Collides with Patent Law." Journal of Legal, Ethical and Regulatory Issues 23, no. 2 (2020): 1-10. HeinOnline.
Ibid.