When Inventee Becomes the Inventor: Can AI be a Patent Inventor?

by Pranita Dhungana

“Why did ChatGPT go to the therapist? Because it had too many layers and couldn’t figure out which one was the true self!” I prompted ChatGPT to tell me a joke about itself, and it responded with the “joke” above. A bit too eerily self-aware for my liking!

Artificial Intelligence (“AI”) has taken the world by storm. The arrival of AI chatbots like ChatGPT has made AI accessible to small businesses that might not otherwise have the resources to develop an AI system in-house. As of 2020, 29% of small and medium businesses had adopted AI. That number must be higher today since AI commonly exists across multiple facets of business, like customer service, marketing, sales, data analysis, inventory management, accounting, and even research & development.

But what happens when your AI system creates an invention? Can it be listed as the inventor on a patent application?

The Answer Depends on the Country You Are In

The United States

The Federal Circuit recently answered this question in Thaler v. Vidal. According to this case, AI cannot be an inventor.

Dr. Thaler developed an AI system called Device for the Autonomous Bootstrapping of Unified Science (“DABUS”). DABUS, without any human involvement, invented a flame device used in search-and-rescue missions and a food/beverage container. Dr. Thaler sought to patent both of these inventions. Interestingly, Dr. Thaler filed both patent applications with DABUS listed as the inventor. The United States Patent and Trademark Office (“PTO”) rejected the applications, reasoning that AI could not be the inventor on a patent. Dr. Thaler unsuccessfully challenged the PTO’s rejection in the U.S. District Court for the Eastern District of Virginia and appealed its decision to the Federal Circuit.

Relying solely on statutory interpretation, the Federal Circuit determined that AI cannot be a patent inventor. The Patent Act (“Act”) defines an inventor as the “individual” who invented the subject matter. The Federal Circuit found the Act’s multiple references to “individual” compelling. Although the Act does not define “individual,” the United States Supreme Court has defined it as a human being or a person. In addition to considering how we use the word individual in everyday use and how dictionaries define it, the court also found support in the Dictionary Act. The Dictionary Act confirms that an individual is a human being and is different from artificial entities like corporations. The Act also uses the personal pronouns “himself” and “herself” to refer to the inventor, as opposed to “itself,” further showing that Congress intended for only humans to have patent inventorship. Finally, the Act requires inventors to submit an oath that they believe themselves to be the original inventor. Since the record was void of any indication that AI can form a belief, and because Dr. Thaler had submitted the oath himself on DABUS’s behalf, the court found no ambiguity in the Patent Act that an inventor must be a natural person.

Regardless of the negative outcome in the U.S., Dr. Thaler has continued his global campaign for the recognition of AI as an inventor.

Europe

Courts across Europe have aligned with the United States in rejecting DABUS’s applications, finding that their patent laws fail to recognize AI as an inventor. The European Union’s European Patent Office specified that under the European Patent Convention, an inventor on a patent application must have “legal capacity,” which is the ability to be the subject of rights and duties. Current laws do not recognize the rights and duties of AI.

Australia

Although a Federal Court of Australia initially ruled that the Australian Patent Act did not limit inventorship to humans, a higher court reversed the ruling based on the finding that the Australian Patent Act confers a patent for human endeavor. Therefore, Australia has also aligned with other jurisdictions globally.

However, an outlier has emerged in South Africa.

South Africa

South Africa remains the only country to have granted a patent to one of DABUS’s inventions. However, commentators have questioned the significance of this grant since South Africa does not have as substantive of a patent examination process as other countries. Specifically, South African patent laws do not define “inventor,” and its patent approval procedure seems to be nothing more than a simple assessment of whether the paperwork was filed correctly.

Regardless, patent rights are territorial, so a patent granted in South Africa is enforceable only in South Africa.

So what should business owners do?

Trade Secret as an Alternative to Patent Protection

Entrepreneurs who implement AI systems in their businesses should understand that their AI-invented inventions are not eligible for patent protection, and there is no way around that bar in most global jurisdictions. In Thaler, the Federal Circuit distinguished DABUS’s inventions from those not entirely made by AI. However, the court provided no guidance on inventions made with the assistance of AI since that was not the issue before the court.

Given this uncertainty, entrepreneurs should utilize trade secrets to protect their AI-invented or AI-assisted inventions. Trade secret protection applies to almost anything that has value because it is not known and is sufficiently secret. In addition to not having a formal registration requirement like patents, trade secret protection also covers those inventions that are not eligible for patent protection, including those that are AI-invented or AI-assisted.

Due to the trade secret secrecy requirement, inventions that are customer-facing or prone to reverse engineering may not qualify. However, for inventions used internally trade secret protection is a viable alternative to patents. To enjoy trade secret protection, business owners should implement measures to maintain secrecy. These measures may include having employees sign confidentiality agreements or NDAs and limiting the distribution of information both inside and outside the business.

This post has been reproduced and updated with the author’s permission. It was originally authored on March 24, 2023 and can be found here.


Pranita Dhungana, at the time of this post, is a recent graduate of Penn State Dickinson Law who is now pursuing Intellectual Property law. She also has a B.S. in Chemistry.

 

 

 

Sources:

ChatGPT, OpenAI (March 21, 2023).

Forbes, AI Stats News: Only 14.6% of Firms Have Deployed AI Capabilities in Production (January 13, 2020) https://www.forbes.com/sites/gilpress/2020/01/13/ai-stats-news-only-146-of-firms-have-deployed-ai-capabilities-in-production/?sh=1e30a55c2650.

Thaler v. Vidal, 43 F.4th 1207 (Fed. Cir. 2022).

Kingsley Egbuonu, The Latest News on the DABUS Patent Case, IPStars (March 17, 2023) https://www.ipstars.com/NewsAndAnalysis/The-latest-news-on-the-DABUS-patent-case/Index/7366.

Andrew J. Gray IV et al., Copyright, Patent, or Trade Secret Protection for AI Content: Challenges and Considerations (February 10, 2023) https://www.morganlewis.com/pubs/2023/02/copyright-patent-or-trade-secret-protection-for-ai-content-challenges-and-considerations#:~:text=Along%20with%20a%20low%20cost,inventions%20made%20by%20AI%20technologies.

Author: Prof Prince

Professor Samantha Prince is an Associate Professor of Lawyering Skills and Entrepreneurship at Penn State Dickinson Law. She has a Master of Laws in Taxation from Georgetown University Law Center, and was a partner in a regional law firm where she handled transactional matters that ranged from an initial public offering to regular representation of a publicly-traded company. Most of her clients were small to medium sized businesses and entrepreneurs, including start-ups. An expert in entrepreneurship law, she established the Penn State Dickinson Law entrepreneurship program, is an advisor for the Entrepreneurship Law Certificate that is available to students, and is the founder and moderator of the Inside Entrepreneurship Law blog.