Inventors Using AI Can Secure IP Protection Via Pannu Factor Test

Artificial intelligence can be a powerful creative tool, but it can also present challenges for inventors seeking intellectual property protections. As more people experiment with AI, the question of ownership has taken on a new importance: When AI is involved, who or what is the real inventor?

The answer has serious ramifications for IP rights. Inventors that rely too heavily on AI tools may find themselves unable to secure patents for their creations because they may not meet the formal criteria for inventorship.

Here’s an overview of where these issues currently stand and how inventors can safely leverage AI tools—without sacrificing IP protections.

Can AI be an inventor? Short answer: No.

However inventive and creative AI tools may appear, they can’t be sole or even joint inventors on a patent under existing US IP law. That right solely belongs to humans, since patent applications and patents themselves can’t list any entity that isn’t a “natural person” as the inventor, even if an AI system played an instrumental role.

Qualifications

To qualify as an inventor, a person needs to have significantly contributed to the invention. So, what constitutes a “significant contribution” for patent purposes?

Aspiring inventors can look to the Pannu factors for answers. Stemming from 1998’s Pannu v. Iolab Corp, these requirements are used to determine whether an individual made significant contributions to an invention as part of a human team, but they apply when using AI as a creative partner too.

Under Pannu, each named inventor for a patent must satisfy all three of these provisions:

• Contribute in some significant manner to the conception or reduction to practice of the invention
• Make a contribution to the claimed invention that is not insignificant in quality, when that contribution is measured against the dimension of the full invention
• Do more than merely explain well-known concepts and/or the current state of the art to the real inventors

In early 2024, the US Patent and Trademark Office issued inventorship guidance specific to AI-assisted inventions, clarifying that while use of AI doesn’t negate a person’s contributions as an inventor, the AI can’t come up with the entire idea.

Recognizing an AI’s output to a general prompt doesn’t make one an inventor, especially if the properties and utility of the output are apparent to anyone.

Applying Pannu

To see how these principles play out, consider this example: Engineers John and Cathy are looking to design a chair that looks like an avocado with help from a generative AI tool.

Given the prompt, “Create an original design for an avocado chair,” the AI tool returns a preliminary design for a chair with a seat bottom and seatback resembling the shape, color, and texture of an avocado, along with a lumbar support that looks like an avocado pit.

There are three different paths John and Cathy might take from here, with different implications under Pannu:

Attempt to patent the AI tool’s chair design as-is. John and Cathy recognized a problem (the lack of an avocado chair on the market), prompted the AI tool to fix this problem, reviewed the AI output, and promptly submitted a patent application.

Recognizing a problem isn’t the same as conceiving an idea, and the prompt they fed into the AI tool restated that problem, with no inventive contribution. What’s more, they reviewed the AI tool’s work, adding nothing in the process.

Verdict: Not inventors. John and Cathy failed to meet the first Pannu factor, since they didn’t contribute anything to the conception of the invention.

Add small details, then attempt to patent it. Using the schematic exactly as outputted by AI, John and Cathy select fabrics and wood to build the chair. They include this additional information in the patent application.

Verdict: Not inventors. On top of still failing to meet the first Pannu factor, John and Cathy also don’t meet the second, since selection of well-known materials is an insignificant contribution compared with the invention as a whole.

Innovate on the AI’s design. From the initial AI design, John designs a removable lumbar support and a mechanism that allows it to be attached to the seatback and bottom in different configurations. Cathy designs an alternative lumbar support that can move vertically along the seatback and lock in different positions.

Verdict: Inventors. In this version, John and Cathy are proper inventors since they satisfy all three Pannu factors. The new lumbar designs are significant contributions to the invention, their experimentation on the general idea resulted in a specific arrangement and design that are integral to the invention, and their contributions are more than just well-known explanations.

As more inventors turn to AI tools in the creative process, they should be careful not to offload too much of the creative work on their AI counterparts. Keeping the Pannu factors and USPTO guidance front-of-mind can help preserve their rights to their creations for years to come.

This article does not necessarily reflect the opinion of Bloomberg Industry Group, Inc., the publisher of Bloomberg Law and Bloomberg Tax, or its owners.

Author Information

Brett Krueger is co-leader of the patent practice group at Honigman and has two decades of experience in patent prosecution.

Grant Griffith is partner at Honigman whose practice focuses on patent prosecutions spanning a wide range of technologies.

Ryan Christiansen is an IP associate at Honigman and former automotive engineer.

Write for Us: Author Guidelines