Authored by Babak Akhlaghi on March 28, 2025. I’ve watched with growing fascination as artificial intelligence transforms nearly every sector of our economy. But as someone deeply immersed in patent law, I’m particularly struck by how AI is forcing us to reexamine fundamental assumptions that have underpinned our innovation ecosystem for centuries.
The day is approaching when we’ll simply identify a problem and ask an AI system for the solution. What happens to patent law then?
AI-assisted inventions currently qualify for patent protection as long as a human makes a “significant contribution” to conception. The USPTO formalized this in February 2024 guidance on AI-assisted inventions, drawing upon existing case law in Thaler v. Vidal, 43 F.4th 1207, 1213 (Fed. Cir. 2022), cert denied, 143 S. Ct. 1783 (2023) and Pannu v. Iolab Corp., 155 F.3d 1344 (Fed. Cir. 1998).
Under this framework, a patent application must name a natural person as an inventor, and that person must have made a “significant contribution” to the claimed invention. The Pannu ruling established a three-part test: the person must (1) contribute significantly to conception, (2) make a contribution not insignificant in quality when measured against the full invention, and (3) do more than merely explain well-known concepts or current state of the art.
Simply identifying a problem and feeding it to an AI system doesn’t meet this threshold. But crafting specific prompts to elicit particular solutions from AI can constitute significant contribution.
This all works well enough today. But what happens when AI evolves toward general intelligence?
What counts as “obvious” is determined by an ordinary person skilled in the art. The recent Federal Circuit decision in AliveCor v. Apple reminded us that this hypothetical person isn’t an automaton—they can reason and draw common-sense motivations even when not explicitly mentioned in prior art. AliveCor, Inc. v. Apple Inc., Case Nos. 2023-1512, 2023-1513, 2023-1514 (Fed. Cir. March 7, 2025) (Hughes, Linn, Stark, JJ.). You can ready my full summary here.
The Court held it would have been obvious to apply machine learning for arrythmia detection given the direction of technology in adopting machine learning in biomedical contexts. This reasoning leads to a crucial insight: a person skilled in the art equipped with powerful AI has substantially more knowledge than one without it.
This means the bar for non-obviousness is rising. And it will continue to rise as AI capabilities advance.
Interestingly, AI advancement might push more companies toward patents rather than trade secrets. With sufficiently advanced AI, reverse engineering solutions long protected by secrecy becomes much easier.
I often use the Coca-Cola formula example in my entrepreneurship class at the University of Maryland. No one has successfully replicated it, making it perhaps the world’s most famous trade secret. But when AI reaches general intelligence and can taste food the way humans do, how long before it replicates the formula or creates alternatives with identical taste profiles?
The traditional protection of trade secrecy may become increasingly vulnerable in an AI-augmented world.
These changes will increase demand for seasoned patent attorneys who can craft claims with broad scope that withstand both design-arounds and patentability challenges. It will become essential to draft independent claims with different scopes to minimize design-around opportunities.
Patent practitioners will need to anticipate various workarounds and capture them in specifications, while maintaining multiple continuation applications to adjust claims when competitors’ AI systems find ways around existing patents.
The field of AI and machine learning patent drafting is becoming highly specialized, much like quantum computing. You need attorneys with relevant technical experience who can provide specific technical details rather than broad functional statements.
At my firm, as we draft more of these applications and observe how examination unfolds, we continuously adapt our techniques to meet future challenges.
I anticipate a bifurcation in the patent world. Incremental innovations will become increasingly difficult to patent as AI-enhanced obviousness standards rise. Meanwhile, truly groundbreaking technologies—like advances in quantum computing—will still receive robust protection.
This may actually benefit the innovation ecosystem by reducing the clutter of marginal improvements while strengthening protection for fundamental advances.
Different jurisdictions will take different approaches to AI inventorship. Companies will naturally seek protection in countries with more permissible laws for patenting AI-generated innovations—but only when there’s market advantage in such protection.
I believe laws will evolve as the interaction between humans and AI continues to develop. We’ll likely see more relaxed standards around AI-generated innovations, or risk falling behind technologically as key innovations migrate to countries offering better protection.
If fewer innovations become patentable due to AI-enhanced obviousness standards, creators will seek alternative protection mechanisms. We might see increased reliance on copyright for software, though copyright faces its own AI challenges. For example, the recent Thaler v. Perlmutter, Case No. 23-5233 (D.C. Cir. Mar. 18, 2025) (Millett, Wilkins, Rogers, JJ.) decision reaffirmed that human authorship is necessary for copyright registration.
AI models rely on vast amounts of training data, raising questions about ownership—who owns the training data may ultimately own AI-generated content. Data ownership for training will become a significant commodity, leading to complex licensing and contractual agreements.
Despite these challenges, patents and copyrights remain vital tools for fueling innovation. Without patents, inventors have less incentive to disclose their inventions to the public. Patents provide the exclusivity needed to commercialize technology in competitive landscapes.
Our IP system must evolve though with technology or risk irrelevance. We may be closer to general AI than many realize, and when it emerges, we must be ready to protect resulting innovations.
The question isn’t whether these innovations should be protectable, but who will own them and under what circumstances. If companies owning expensive general AI systems can claim all innovations generated by their machines, how will smaller companies with limited budgets compete?
We may need to shift toward more open innovation models for purely AI-generated innovations to prevent monopolization by a few wealthy companies. A hybrid approach could emerge—strong patent protection for human-AI collaborative inventions alongside open ecosystems for purely AI-generated innovations.
The challenges ahead are significant, but so are the opportunities. Those who understand the evolving relationship between AI and intellectual property will be positioned to navigate this changing landscape effectively. The rules are being rewritten, and we all have a stake in how they develop.
Want more valuable tips on navigating patent law and protecting your intellectual property? Join our free weekly newsletter for the latest updates, expert advice, and exclusive content straight to your inbox. Subscribe today and never miss out on essential information that can make all the difference in your patent journey! Subscribe to our Newsletter