Inventors are increasingly using generative artificial intelligence (AI), including large language models, and other AI systems, to brainstorm ideas, and refine solutions to engineering and product design problems which raises important questions about patent protection.
If AI helped generate your invention, can you still patent it? In many cases, the answer is yes, provided that one or more human inventors made a meaningful contribution to the conception of the claimed invention.
Relying on AI-developed contributions
Patent systems around the world generally require that an inventor be a natural person. AI systems cannot be listed as inventors. However, using AI as a tool during the inventive process is permissible, just as inventors have long relied on software, simulation tools, design tools, and research databases to assist in developing inventions. AI can assist throughout the inventive process, but the critical requirement is that at least one human inventor contributes to the conception of the claimed invention.
Your contribution vs. the AI’s contribution
Inventorship generally turns on conception, meaning the formation of a definite and workable idea of the invention. If AI played a role in developing your invention, you should be able to identify and explain what you contributed to the inventive concept, as distinct from what the AI generated.
Meaningful human contribution may include:
- Defining the problem, evaluating and refining AI-generated outputs;
- Testing, modifying or validating proposed solutions using your judgment; and
- Combining multiple concepts into a new workable solution.
Simply presenting a general problem to an AI system and accepting whatever it produces, without meaningful intellectual input or refinement, may make it difficult to establish a valid inventorship position. The more you can point to your own reasoning, modifications, and technical judgment, the stronger your position is likely to be.
Risk of multiple people arriving at similar solutions
Because generative AI tools are often trained on overlapping sources of information and may produce similar outputs in response to similar prompts, it is possible that different users could arrive at similar solutions.
This creates several practical risks. Similar inventions may be independently developed by multiple inventors, resulting in competing patent filings. It may also become more difficult to establish that a particular solution is novel and distinct from what others have already developed, and whether the resulting invention would have been obvious to a person skilled in the relevant field.
Meaningful contribution matters especially if challenged
If your patent is ever challenged, whether during litigation, or otherwise, you may need to explain and justify your inventive contribution. For this reason, it is important to document your efforts throughout the inventive process rather than merely preserving the final AI-generated output. Consider maintaining records of:
- The problem you set out to solve;
- The prompt iterations you developed;
- Technical assumptions you made;
- AI-generated outputs that were accepted, modified, or rejected;
- Why certain approaches were discarded;
- Experiments, testing, and validation efforts; and
- Your own modifications, refinements, and contributions.
Patent validity and inventorship disputes can arise many years after a patent application is filed. Detailed records created during the development process provide valuable evidence and are often more persuasive than recollections formed after the fact.
Ensure the solution is accurate and actually works
Generative AI systems can produce solutions that appear convincing but are technically flawed, impractical, or simply incorrect. Before relying on an AI-generated idea or filing a patent application, independently verify that the proposed solution is technically sound, workable, and accurately reflects what you intend to invent. You should be able to explain why the invention works, not merely that an AI system suggested it.
Patent applications generally must describe the invention in sufficient detail to enable a person skilled in the field to make and use the invention. An unverified AI output is therefore a weak foundation for both innovation and patent protection.
Don’t Assume AI found all Prior Art
Many inventors are using AI tools to assess whether an invention may be patentable or to identify potentially relevant prior art. While such tools can provide useful preliminary insights, they should not be treated as a substitute for a comprehensive prior art search.
AI systems may:
- miss relevant public disclosures;
- generate inaccurate citations;
- misinterpret technical subject matter; or
- provide overly optimistic assessments regarding patentability.
Protect your confidentiality
Before entering invention details into any AI platform, carefully review the provider’s terms of use, data-handling practices, and privacy-confidentiality protections.
Some AI platforms may retain user inputs, use them for system improvement or training purposes, or permit data access under certain circumstances. Disclosure of sensitive invention details could compromise trade secret protection or create unintended risks for future patent filings.
Stay current
The legal treatment of AI-assisted inventions continues to evolve around the world. Patent offices are actively considering questions relating to inventorship, ownership, disclosure obligations, and the role of AI in innovation. Inventors should consult qualified patent professionals regarding current requirements in the jurisdictions where protection is sought.
Guidance specific to Canada
Canadian patent law is governed by the Patent Act and administered by the Canadian Intellectual Property Office (CIPO). The following points are particularly relevant for seeking patent protection in Canada:
- Under current Canadian law and practice, an AI system cannot presently be named as an inventor on a Canadian patent application.
- Inventors should be prepared to identify and document their contribution to the inventive concept.
- Canadian patents must satisfy the traditional patentability requirements, including novelty, non-obviousness, utility, and adequate disclosure. The use of AI during the inventive process does not alter these requirements.
- Canada does not currently have specific legislation governing AI inventorship or a general requirement to disclose the use of AI tools in preparing or developing an invention described in a patent application.
Conclusion
Generative AI can be a powerful aid in innovation, helping inventors explore ideas, analyze technical problems, and accelerate aspects of the inventive process. However, AI should be viewed as a tool rather than a substitute for human ingenuity.
Inventors who use AI should focus on making meaningful contributions of their own, independently validating proposed solutions, maintaining clear records of their work, protecting confidential information, and seeking professional guidance where appropriate.
Disclaimer: This publication is provided for general informational purposes only and does not constitute legal advice. Inventors should consult a qualified patent professional regarding their specific circumstances.
About the author
Kushal Shah is a Patent Lawyer and registered Patent Agent at Perley-Robertson. He advises innovators, entrepreneurs, research institutions, and businesses on protecting, managing, and commercializing intellectual property. His practice includes patent drafting and prosecution, intellectual property strategy, portfolio development, licensing, and technology commercialization. Kushal works across pharmaceuticals, life sciences, biomedical devices, software, artificial intelligence, and other high-technology fields. He
helps clients develop intellectual property strategies that support their innovation goals and broader commercial priorities.
