AI has moved from science-fiction speculation, like the 2001 film “A.I. Artificial Intelligence” by Steven Spielberg, to a real technology accessible to nearly everyone. E-commerce recommendation engines, voice-activated digital assistants, email spam filters, wayfinding apps, and workplace efficiency tools are all powered by AI. As debates around whether AI is poised to replace people’s jobs and whether AI is can ever be safe continue, one big question remains: who owns AI? Patent law is still playing catch-up, according to Michael Carey, an AI-focused patent attorney at F. Chau & Associates.
“Typically, it takes a while for the value of various types of patents to become clear,” said Carey. “But innovative companies don’t want to wait. So, companies who are on the cutting edge of developing technology typically spend a portion of their budget getting patent protection.”
Who Owns AI?
The question of who owns AI is growing alongside the technology’s proliferation. And the short answer is, “it depends.” No one owns the actual idea and concept of artificial intelligence, machine learning, or large language models. Companies own specific implementations featuring trained models, infrastructure, and proprietary data. For instance, OpenAI owns ChatGPT, Google owns Gemini, and Anthropic owns Claude.
Where things can get murky is in who owns the actual output of what a model or other AI tool produces. A company designing a product using an LLM as a launching pad or a designer building a new logo that’s largely inspired by an AI creation may not have strong claims over ownership.
What is the Current Legal Standard for AI Patent Ownership?
In February 2024, the U.S. Patent and Trademark Office (USPTO) issued Inventorship Guidance for AI-Assisted Inventions. In short, USPTO stated that AI-assisted inventions could receive a patent if one natural person made a significant contribution to the conception that didn’t only explain well-known concepts or current knowledge.
The guidance also said certain tools, including AI systems, could perform acts that “could constitute inventorship under our laws” if those acts were performed by a human. The AI could contribute so greatly that it would be viewed as a co-inventor.
Less than two years later, USPTO rescinded that initial guidance. It released the Revised Inventorship Guidance for AI-Assisted Inventions in November 2025. The biggest change: AI can no longer be a potential co-inventor. Instead, it’s a “tool that assists,” much like “laboratory equipment, computer software, research databases,” or any other number of current physical or technological pieces that aid humans.
USPTO claims that while patent law “currently requires a human inventor and does not allow patenting of inventions made solely by AI, patents can be granted on at least some inventions that human inventors make with AI assistance.”
The ambiguity could lead to patent lawsuits, though companies may be hesitant to go down that path.
“When they do happen, they’re really huge,” Carey said. “I think that where the industry is at right now, people are almost getting ‘defensive’ patents.”
“In case they need to go to war with one of their big rivals, they can dig out all their patents and then try and have some ammunition. It’s almost like mutually assured destruction. Like: ‘Don’t sue me, because if you do, we’re going to countersue,’ and we’ve got this big patent portfolio to do that.”
The motivation behind AI patents is essentially the same for any other patents, Carey said. Namely, they allow inventors to capture more value from their creations and therefore provide an additional incentive for companies to invest in innovation.
What Does the USPTO’s Revised Guidance Require in Practice?
The U.S. Congress itself notes that the 2025 Inventorship Guidance “arguably provides less clarity in the event that AI is able to supply some or all of the core idea for an invention.” However, the relative newness of artificial intelligence patent law, combined with the somewhat limited visibility into most AI-powered technologies, makes patent enforceability something of an open question.
“The AI patent world is still a little immature,” Carey said. “People haven’t found the patterns or the easily exploitable ways to enforce AI patents. And there’s a bit of a challenge in enforcing AI patents because some of the key elements of AI patenting, including the architecture and training, are actually a bit harder to observe for AI than even with other software tools.”
“If you are coming from a traditional software background, you’re used to having code that carries out a specific function. With AI patents, it’s not the same. You have millions of parameters, and they get learned by the algorithm, and it’s kind of a black box.”
In practice, companies can take proactive steps to avoid potential AI pitfalls or at least minimize the challenges they pose. Documenting human contributions to AI tools, such as prompting, configurations, inputs, and choosing (or iterating on) AI outputs, can point to humans having a hand in inventing a product.
USPTO might reject a patent as being invalid if it lists AI as an inventor. Disclosure forms, internal policies, and pending applications should all receive a thorough audit to ensure only humans are named as inventors. Again, in these situations, their contributions should be documented to showcase their role in the invention.
However, AI is increasingly discovering new drugs and materials and conceiving novel designs. We may soon see new guidance on the intellectual property that artificial intelligence generates.
How Are Patent Attorneys and IP Teams Using AI in Their Own Practice?
As artificial intelligence continues to develop, the technology can serve as both a creation and enforcement tool. AI has the potential to devise inventions that are worthy of patent protection and it can also help companies and its lawyers more easily identify when protected assets have been infringed upon.
Patent attorneys and IP teams are using AI in their own practice, with the modular nature of the work lending itself to several areas of assistance. For example, a patent practitioner often receives patents that are too broad, too narrow, or written by an engineer or marketer, all of which add challenges. AI can turn that raw information into something more structured, identifying the inventive concept, technical advantages, gaps in the data, and questions that still need to be answered.
AI also lends a hand in applications like competitive intelligence, highlighting what filing areas are overcrowded, where there may be gaps, and what competitors are doing in the product market. And rather than using AI to draft claims, strategic patent attorneys are having it act as an examiner or defendant to poke holes, such as disputes in claim construction or infringement issues. That quality check simulates how a human review board might act and enables proactive fixes.
Brent Nix, Principal at Johnson, Marcou, Isaacs & Nix, LLC, considers AI to be more of a “sparring partner” than a wise guide people should follow without question.
“AI is excellent for bouncing ideas around and stress-testing claim language,” he wrote. “But it won’t reliably lead you to the right prosecution strategy…It’s like having a very junior patent attorney who’s read every patent but has questionable strategic judgment—great for brainstorming claim language, terrible for final prosecution decisions.”
How to Build an Effective Patent Strategy for AI-Assisted Inventions
Carey said that it is a mistake to draft patents for artificial intelligence the same way one would draft a patent for any other software application.
For instance, an ineffectual AI patent might use the algorithm itself, which is often powered by cloud computing, as the basis for several different claims, and only alter the language to protect a different “statutory class” in each claim.
While this sort of patent structure is simple for attorneys to follow, Carey said it fails to provide the broadest possible protection.
Instead, Carey suggested that attorneys should draft AI patents with three distinct types of claims: application, architecture and training. An application claim will explain what the invention does and how it does it, describing the inputs for the AI invention, the function of the system, and how the tool is applied to solve a problem.
A good architecture claim, Carey said, will hit a “sweet spot” by providing a description of architectural components that is neither too broad nor too narrow. Lastly, a training claim will explain how tools collect data, predict results, and then update model parameters over time.
These sorts of claims address what is unique about AI tools, and also offer multiple ways to describe an inventive concept, potentially increasing enforceability. Carey added that organizations need to work with attorneys who have experience in artificial intelligence patent law.
Experienced attorneys have a better grasp on an effective structure for AI patents. They also understand the subject-specific content and terminology necessary to craft a strong patent that will be accepted by the U.S. Patent and Trademark Office and withstand challenges in court.
Deep domain expertise also “allows me to quickly spot when AI is helping versus when it’s going down the wrong path,” according to Nix. “Without that foundation, you’re flying blind. I can immediately recognize when an AI-generated claim set misses critical prior art distinctions or when a proposed claim amendment inadvertently narrows scope beyond what’s necessary to overcome a rejection.”
“It is a little bit harder to detect infringement in AI patents,” Carey said. “But in my view, it’s actually a little easier to get the patent in the first place, if you draft it correctly.”
That will give companies yet another tool on the innovation battlefield.
Related:
This is an updated version of the article originally published on January 22, 2022.
Calvin Hennick is a contributing writer. His work appears in BizTech, Engineering Inc., The Boston Globe Magazine and elsewhere. He is also the author of Once More to the Rodeo: A Memoir. Follow him on Twitter @CalvinHennick.
Chase Guttman contributed to this update of this article. He is a technology writer. He’s also an award-winning travel photographer, Emmy-winning drone cinematographer, author, lecturer and instructor. His book, The Handbook of Drone Photography, was one of the first written on the topic and received critical acclaim. Find him at chaseguttman.com or @chaseguttman.
© 2026 Nutanix, Inc. All rights reserved. For additional information and important legal disclaimers, please go here.
Related Articles
