Artificial intelligence is transforming how inventions are conceived, developed, and commercialised. From drug discovery to materials science to software engineering, AI systems are increasingly integral to research and development processes. Yet as these technologies proliferate, they collide with a patent system designed for human inventors, raising fundamental questions about who (or what) can be an inventor, what subject matter is patentable, and how businesses should navigate the emerging legal landscape.
In 2025, Canadian patent law is grappling with these questions in real time. The Patent Appeal Board has ruled that AI cannot be named as an inventor. Computer-implemented inventions face a complex eligibility framework. Courts have imposed sanctions for undisclosed use of generative AI in litigation. For technology companies and innovators, understanding these developments is essential, and requires guidance from seasoned Intellectual Property Lawyers in Toronto with expertise in patent prosecution, enforcement, and litigation strategy.
This article examines the current state of AI and patent law in Canada, including the landmark DABUS decision, the evolving framework for computer-implemented inventions, and the procedural obligations that now apply when AI is used in litigation. For businesses deploying AI in their innovation processes, understanding these rules is no longer optional, it is a competitive necessity.
AI Cannot Be an Inventor in Canada: The DABUS Decision
On June 5, 2025, the Commissioner of Patents refused Canadian Patent Application No. 3,137,161 following a recommendation by the Patent Appeal Board (PAB). The decision, reported as Thaler, Stephen L. (Re), 2025 CACP 8, represents Canada’s first, and likely definitive, ruling on whether an AI system can be named as an inventor under Canadian patent law.
The DABUS Application
The patent application at issue related to a fractal-based food container and a flashing emergency beacon. What made it unique was not the inventions themselves, but the identity of the listed inventor: DABUS (Device for the Autonomous Bootstrapping of Unified Sentience), an AI system created by Dr. Stephen Thaler.
Dr. Thaler had filed similar applications in multiple jurisdictions worldwide, naming DABUS as the sole inventor and himself as the inventor’s legal representative. His explicit goal was to challenge and expand the legal definition of “inventorship” under patent law. Courts and patent offices in the United States, United Kingdom, European Patent Office, and Australia had all rejected the proposition that an AI system can be recognised as a named inventor.
The Patent Appeal Board’s Reasoning
The PAB’s analysis rested on statutory interpretation. Although “inventor” is not expressly defined in the Patent Act or the Patent Rules, the Board concluded that the ordinary meaning of the term refers to a “person”, specifically, a natural person.
The Board identified multiple provisions in the Patent Act that connote human characteristics. Section 31, for example, speaks of an inventor who “refuses to proceed” or whose “whereabouts cannot be ascertained”, concepts incompatible with an AI system. The Patent Rules require inventors to have names and postal addresses, requirements that, the Board noted, “weren’t contemplated as applying to software or AI.”
The Board also rejected Dr. Thaler’s policy arguments. He had invoked the ancient Roman principle of “accession” (the owner of property owns its fruits) to claim ownership of DABUS’s inventions. The Board found no precedent for applying accession to intangible intellectual property. Dr. Thaler’s warning that denying AI inventorship would “chill innovation” was similarly rejected; the Board noted that pharmaceutical companies have obtained thousands of patents using AI as a tool in development, suggesting no barriers to innovation exist under the current framework.
The Appeal to Federal Court
Dr. Thaler filed a Notice of Appeal to the Federal Court on December 5, 2025. The appeal challenges the legal conclusion that “inventor” must be limited to natural persons, arguing that interpretation of the Patent Act should evolve to reflect new technological realities, including machines that can autonomously generate inventions.
The appeal is expected to be heard in the second half of 2026. While the PAB decision is not binding law, it will likely guide patent examiners in Canada until a court rules otherwise.
AI-Assisted Inventions: What Can Be Patented
The DABUS decision addresses a narrow question: can an AI system be named as an inventor? It does not preclude patenting inventions in which AI played a role.
The “Significant Human Contribution” Standard
Under Canadian inventorship doctrine, a human must “contribute to the substance of the invention” to qualify as an inventor. Merely being involved in verification is insufficient; presenting a problem to another is not an act of invention. These principles apply equally when the “other” is an AI system.
This aligns with the approach taken by the United States Patent and Trademark Office (USPTO), which issued Inventorship Guidance for AI-Assisted Inventions in February 2024. The USPTO guidance provides that AI-assisted inventions are not categorically unpatentable. If a human makes a “significant contribution” to the inventive concept, such as designing the AI workflow, defining the problem, or interpreting the model’s outputs, the application may proceed with that human as inventor.
While Canada has not issued equivalent formal guidance, the principles are consistent with Canadian Intellectual Property Office (CIPO) practice. The patentability inquiry asks whether the claimed invention meets the requirements of section 2 of the Patent Act: whether it is a “new and useful art, process, machine, manufacture or composition of matter.” The question is not whether the inventor used AI, but whether the claimed process represents a concrete, reproducible, and useful application of technology.
Practical Considerations for Patent Applications
For businesses using AI in their R&D processes, several practical considerations emerge:
Document human contributions. Maintain records of how human inventors contributed to the inventive concept. including problem definition, experimental design, data interpretation, and refinement of AI outputs. This documentation may be critical if inventorship is ever challenged.
Ensure enablement and sufficiency. The “black box” nature of many AI systems can create challenges for patent specifications. If the person preparing the specification cannot explain how the invention was derived or how it is performed, it may not be possible to make a full disclosure enabling others to carry out the invention. Ensure that claims can be supported by a description that meets the disclosure requirements of section 27 of the Patent Act.
Consider disclosure obligations. While there is currently no express requirement in Canada to disclose the use of AI in the inventive process, evolving international standards, including the USPTO guidance, suggest that best practices include transparency about AI involvement.
Patentable Subject Matter for Computer-Implemented Inventions
Beyond inventorship, businesses seeking to patent AI-related technologies face a separate challenge: demonstrating that the claimed subject matter is patentable under Canadian law.
The Current Framework: PN2020-04 and MOPOP
CIPO’s current approach to computer-implemented inventions is set out in Practice Notice PN2020-04 and the Manual of Patent Office Practice (MOPOP). The framework emerged from litigation in Choueifaty v. Canada (Attorney General), 2020 FC 837, and subsequent decisions.
Under the current guidance, subject matter eligibility is generally recognised if: (1) the claimed invention results in improved functioning of the computer; or (2) the claims recite non-generic computer hardware (such as sensors or other hardware extrinsic to the computer itself). Where software does not improve computer functionality, CIPO will generally not consider a claim to such software to be patentable.
The key goal of the subject matter eligibility analysis is to identify whether the proposed invention “provides a novel and inventive technological solution to a technological problem.” Abstract algorithms, mere scientific principles, and “professional skills” remain unpatentable.
The Benjamin Moore Decisions
The evolution of Canada’s approach to computer-implemented inventions has been shaped by the Benjamin Moore litigation. The case involved patent applications for computer-implemented paint colour selection methods using experimentally derived relationships for colour harmony and colour emotion.
In 2022, the Federal Court proposed a new three-part framework for assessing eligibility. However, the Federal Court of Appeal overturned this approach in Canada (Attorney General) v. Benjamin Moore & Co., 2023 FCA 168, finding it premature. The FCA directed the Commissioner to redetermine the applications using CIPO’s current guidance (PN2020-04) with the benefit of the FCA’s reasons.
In August 2025, the Commissioner issued redetermination decisions in Benjamin Moore & Co. (Re), 2025 CACP 9 and 2025 CACP 10. All original claims in both applications were rejected on patentable subject matter grounds. The Commissioner found that the claims were akin to a “mere scientific principle or abstract theorem” and comprised unpatentable “professional skills” and “fine arts.” Amended claim sets were also rejected for failing to remedy the defects.
Benjamin Moore has six months from the date of the decisions to appeal to the Federal Court. The outcome could provide further clarity on the boundaries of patentable subject matter for AI and software-related inventions in Canada.
Implications for AI Technologies
While AI cannot be an inventor in Canada, AI itself can comprise patentable subject matter, provided it is not an abstract idea and meets the “physicality” requirement. An AI system that improves computer functionality, or that is implemented in combination with non-generic hardware, may be eligible for patent protection.
The Federal Court of Appeal cautioned in Benjamin Moore FCA that the Commissioner “should keep an open mind and not hastily conclude that the subject matter claimed is not patentable simply because it involves the use of conventional computer technology.” The determination remains “highly fact specific.”
AI Disclosure Obligations in Federal Court Litigation
For patent litigators and in-house counsel, there is a separate dimension to AI and intellectual property: the use of AI tools in litigation itself.
The Federal Court’s May 2024 Notice
On May 7, 2024, the Federal Court of Canada issued a “Notice to the Parties and the Profession” regarding the use of artificial intelligence in court proceedings. The notice was subsequently incorporated into the Amended Consolidated Practice Guidelines on June 20, 2025.
The Notice requires parties to inform the Court and each other if documents submitted to the Federal Court for the purpose of litigation include content created or generated by AI. This must be done through a declaration in the first paragraph stating that AI was used in preparing the document, either in its entirety or for specifically identified paragraphs.
The declaration is required when AI “generates the content itself”, that is, when AI plays a role resembling that of a co-author. It is not required if AI was used merely to suggest changes, provide recommendations, or critique content already created by a human.
Consequences for Non-Compliance
The June 2025 Practice Guidelines warn that “[f]ailure to comply with the Notice may result in consequences for parties and/or counsel, including the imposition of an adverse cost award or an order to show cause why the party or counsel in question should not be held in contempt.”
These consequences have already manifested in case law. In Hussein v. Canada, 2025 FC 1060, the Federal Court criticised the use of AI-generated “hallucinations” in submissions from the applicant’s counsel. The AI tool had fabricated two non-existent cases and “hallucinated” an erroneous legal test.
Associate Judge Moore stated that “[t]he use of generative artificial intelligence is increasingly common and a perfectly valid tool for counsel to use; however, in this Court, its use must be declared and as a matter of both practice, good sense and professionalism, its output must be verified by a human.” The Court found that counsel’s failure to disclose reliance on AI until after four court directions “amounts to an attempt to mislead the Court.”
In a follow-up decision (Hussein v. Canada, 2025 FC 1138), Associate Judge Moore ordered costs personally against the applicant’s counsel in the amount of $100. While modest, the award signals judicial willingness to sanction counsel for AI-related misconduct.
Best Practices for Legal Professionals
The Federal Court’s approach reflects broader trends across Canadian jurisdictions. Ontario has introduced certification requirements for legal authorities and expert reports effective December 1, 2024. Nova Scotia courts require disclosure of AI use, including specification of which AI tool was used.
For counsel using AI tools in litigation, the following practices are essential:
Verify all AI outputs. Before submitting any document to court, confirm that all case citations exist and accurately support the propositions for which they are cited. AI “hallucinations” remain a significant risk.
Disclose AI use. Include the required declaration when AI has generated content in court filings. Transparency is required; concealment may be characterised as an attempt to mislead.
Consider confidentiality. The Law Society of Ontario has cautioned about confidentiality risks when inputting privileged or confidential information into AI tools. Some providers may use this data for training or storage.
What This Means for Your Business
For technology companies and innovators, the intersection of AI and patent law presents both challenges and opportunities.
For R&D-Intensive Businesses
Integrate IP strategy into AI workflows. When deploying AI in research and development, consider from the outset how inventions will be documented, who will be named as inventors, and how patent applications will be drafted to satisfy disclosure and enablement requirements.
Train technical staff on inventorship. Ensure that scientists, engineers, and data scientists understand what constitutes a “significant contribution” to an invention. This understanding will help identify the correct inventors and avoid challenges during prosecution or litigation.
For Software and AI Companies
Draft claims strategically. Computer-implemented inventions face a rigorous eligibility analysis. Work with patent counsel to ensure claims emphasise technical improvements to computer functionality or include non-generic hardware elements where appropriate.
Monitor evolving guidance. CIPO’s approach to patentable subject matter continues to evolve. The outcome of the Benjamin Moore appeals and the Thaler Federal Court appeal could reshape the boundaries of what is patentable.
For In-House Counsel
Establish AI governance policies. If your organisation uses AI tools in legal work, establish clear policies on disclosure, verification, and confidentiality. The reputational and professional risks of AI-related misconduct in litigation are significant.
Coordinate with outside counsel. Ensure that litigation counsel are aware of and complying with court-specific AI disclosure requirements. Different courts have adopted different approaches; Federal Court filings require declarations that may not be required in provincial courts.
Conclusion
The Canadian patent system is adapting, slowly, to the realities of artificial intelligence. The DABUS decision confirms that, for now, only humans can be inventors in Canada. The Benjamin Moore line of cases continues to define the boundaries of patentable subject matter for computer-implemented inventions. And the Federal Court’s AI disclosure requirements impose new obligations on litigators.
For businesses at the forefront of AI innovation, these developments are not merely academic. They affect patent strategy, R&D documentation, litigation risk, and competitive positioning. Early engagement with experienced patent counsel, both for prosecution and litigation, is essential to navigating this evolving landscape.
The questions raised by AI and patent law are unlikely to be fully resolved by courts or patent offices anytime soon. But businesses that understand the current rules, monitor legal developments, and adapt their practices accordingly will be best positioned to protect their innovations in an AI-driven economy.
