Our Blog

CIPO Is Updating Its Software Patent Guidance: What It Means for AI and Quantum Inventions

November 18, 2026

The Canadian Intellectual Property Office has opened a consultation on proposed updates to the Manual of Patent Office Practice (MOPOP), including a substantial revision to Chapter 22 on computer-implemented inventions.

The proposed changes are intended to incorporate CIPO’s March 2026 Practice Notice on patentable subject matter and update the guidance examiners use when assessing software-related inventions. The consultation is open until December 18, 2026.

For companies developing AI, quantum, and other software-driven technologies, the draft guidance provides a useful indication of how CIPO may examine these inventions going forward.

Computer Implementation Alone Is Not Enough

The draft Chapter 22 makes clear that simply reciting a computer will not necessarily make an otherwise abstract algorithm patentable.

Instead, CIPO focuses on whether the claimed subject matter has the required physicality and whether the computer and algorithm together form a patentable practical application.

One particularly important theme is improved computer operation. The draft guidance distinguishes between an algorithm that merely produces a better result and one that improves how the computer performs the task. CIPO identifies examples such as reduced memory usage and fewer processor cycles.

That distinction can be especially important for AI inventions.

AI Inventions Remain Highly Fact-Specific

The draft Chapter 22 includes specific guidance for AI and machine-learning inventions.

CIPO proposes treating AI-related inventions like other computer-implemented algorithms. The analysis therefore turns on what the claimed invention actually does and how the computer participates in the invention.

For applicants, the practical takeaway is that technical improvements should be explained clearly in the patent application.

If an AI technique reduces memory requirements, improves processing efficiency, changes how computing resources are allocated, or otherwise improves computer operation, those effects should be described in the specification rather than left for later argument during prosecution.

The same applies to enablement. The draft guidance notes that machine-learning inventions must disclose enough about the training methodology or training data to allow the skilled person to make and use the invention without undue experimentation.

What About Quantum Computing?

The draft Chapter 22 is also notable for expressly recognizing that a processor can include a quantum processing unit.

That is significant because many quantum-related inventions include substantial software or algorithmic components.

For quantum-related inventions, the same general question will arise: does the claimed implementation improve the operation of the computing system or otherwise provide the required physicality?

The draft’s express inclusion of quantum processing units confirms that CIPO’s computer-implemented invention framework is intended to encompass quantum computing systems.

The Practical Takeaway

CIPO’s proposed revisions do not create an entirely new framework. They largely formalize the direction set out in the March 2026 Practice Notice, which superseded older MOPOP references to a “technological solution to a technological problem” and the former problem-and-solution approach.

What the draft guidance does provide is a clearer picture of how CIPO may apply that framework to computer-implemented inventions.

For companies developing AI, quantum, and other advanced computing technologies, the practical lesson is straightforward: technical improvements should be identified and described when the application is drafted, not reconstructed years later during prosecution.

As CIPO continues to refine its guidance, applicants that clearly explain how their inventions improve computer operation, interact with physical systems, or use computing resources in a non-conventional way will be better positioned to navigate Canadian subject-matter eligibility.


Archive