Estimated reading time: 6 minutes
Summary
You built something amazing using AI. It feels revolutionary. Your first instinct is to protect it, to patent it before anyone else can. But a critical question looms: Can AI patent an invention? Or, more importantly, can you patent an invention that AI helped you create?
Key Takeaways
- AI cannot patent inventions; inventors must be human, as clarified in the Thaler v. Vidal case.
- Using an AI tool does not eliminate your chance of receiving a patent if you demonstrate a significant human contribution.
- A significant human contribution can include constructing effective prompts, designing or training the AI, and modifying the output from AI.
- Documenting your contributions from the start is crucial for patenting your AI-assisted invention.
- Ultimately, you can patent your AI-assisted invention if you provide clear evidence of your human involvement.
Table of contents
Prefer to watch instead? Check out the video below.
As a patent attorney, I get calls about this every single week. The excitement from startup founders is contagious. But so is the potential for a devastating legal roadblock.
Let me share a story from a recent call that perfectly illustrates the problem.
A Founder’s Panic
The phone rang. It was a startup founder, his voice buzzing with energy. He told me they had built an incredible new platform using AI and needed a patent immediately.
As he described the technology, I heard about the fantastic, complex outputs the AI was generating. But I heard very little about what the founder and his team actually did. The line between human ingenuity and artificial intelligence was blurry.
I had to ask a tough question. “It sounds powerful,” I said. “But tell me, did the AI do most of the ‘inventing’ here?”
He explained they used an AI tool during development. I pressed him. How they used that tool would determine everything.
I told him bluntly, “If the AI invented this, you can’t get a patent. An AI cannot be named as an inventor.”
Silence.
The excitement in his voice was replaced by a sense of panic. “What do you mean I’m not the inventor?” he finally asked. “I built this!”
This founder had stumbled upon one of the most critical legal questions for today’s innovators. The good news? We now have a clear roadmap from the courts and the U.S. Patent and Trademark Office (USPTO).
The Legal Foundation – An Inventor Must Be Human
The first thing to understand is a landmark court case: Thaler v. Vidal. Dr. Stephen Thaler argued that his AI system, DABUS, had independently created two inventions. He tried to name DABUS as the inventor on the patent applications.
The case went all the way to the Federal Circuit court. The court’s decision was clear and simple. The U.S. Patent Act refers to inventors as “individuals.” The court interpreted “individual” to mean a “natural person” or a human being.
Therefore, an AI system cannot be named as an inventor on a patent. Full stop.
But the court left a huge question unanswered. What happens when a human uses AI as a tool to create an invention? That’s where new guidance comes in.
Guidance for AI-Assisted Inventions
In February 2024, the USPTO released its official Inventorship Guidance for AI-Assisted Inventions. This document, published in the Federal Register, gives patent examiners a framework for analyzing these exact situations.
The best news for innovators is that using AI does not kill your chances of getting a patent.
Instead, the entire analysis focuses on one thing: the significance of the human contribution.
What Is a “Significant Human Contribution”?
To be named an inventor, you must have made a “significant contribution” to the conception of the invention. This is the heart of the matter. Simply pushing a button and letting an AI work its magic is not enough.
So, what counts as a “significant contribution” when using AI? Let’s break it down.
1. Constructing the Prompt
Just presenting a general problem to an AI is not inventive. You can’t simply ask an AI to “solve world hunger” and then claim inventorship over its proposed solution.
However, your contribution may be significant if you construct a specific prompt. A well-crafted prompt can guide the AI toward a solution that embodies your own inventive concept. You are using the AI as a sophisticated tool to execute your vision. Think of it as the difference between asking a person to “paint a picture” versus giving them detailed instructions on the subject, style, color palette, and composition.
2. Designing, Building, or Training the AI
Your contribution is almost certainly significant if you designed, built, or trained the AI to solve a specific problem.
Did you create a unique machine learning model? Did you curate a specialized data set to train an existing model for a new purpose? These actions show that you engineered the tool itself to achieve a particular inventive outcome. This work is a clear contribution to any invention the AI subsequently helps produce.
3. Modifying the Output
Recognizing that an AI’s output is useful is not enough to make you an inventor. You must do more.
A significant contribution occurs when you take the AI’s output and substantially modify it. Perhaps you combine the AI’s output with other components. Maybe you use the output as a starting point but conduct your own experiments to refine it into a final, working invention. This act of refinement, selection, and modification is a key part of the inventive process.
In short, an invention generated purely by an AI, with no meaningful human guidance or refinement, likely cannot be patented. It lacks the required human inventorship.
You can find more resources and information on the USPTO’s website about artificial intelligence initiatives.
A Practical Framework for AI Innovators
By the end of my call with the startup founder, his panic was gone. He felt empowered. He now had a clear framework to analyze his team’s work and move forward.
This framework should change how you approach innovation with AI. You must be intentional and document your contributions from day one.
Here are some actionable steps you can take:
- Document Your Conception: Before you even use an AI tool, write down your initial idea. What problem are you trying to solve? What is your proposed solution?
- Keep Meticulous Logs: Save your prompts. Document why you structured them in a certain way. Keep records of the different outputs the AI generated.
- Record Your Refinements: Make notes on why you selected a particular AI output. Detail every modification you made to that output. Explain how you combined it with other elements to create the final invention.
- Focus on the “Why”: Why did you choose a specific AI model? Why did you train it with a particular data set? Your reasoning demonstrates your intellectual contribution.
AI is Your Tool, Not Your Co-Inventor
The question “Can AI patent an invention?” has a simple answer: No.
But the real question for you is, “Can I get a patent for my AI-assisted invention?” The answer to that is a resounding “Yes,” provided you can demonstrate your significant human contribution.
The law is not trying to stop innovation. It is trying to ensure that patents are awarded to human inventors. By understanding the rules and diligently documenting your unique input, you can confidently use AI to create groundbreaking, patentable technologies.
If you are building with AI and need to protect your core business, it’s critical to get this right. Thinking about inventorship from the start can prevent a near-disaster later. Let’s talk about how your human contribution can stand out where it matters most.
Get Your AI Inventorship Checklist here!
You may also want to read:
