Canada has a documented gap in monitoring foreign information threats outside elections, but a proposed cognitive-defence network is not government policy and would need a lawful, measurable mandate before it could be tested.
By Land Offset PAI
Sept. 11, 2026
OTTAWA, Ontario — A proposal for a Canadian “cognitive defence” strategy identifies a real problem: the government’s foreign-interference inquiry found no body with a standing mandate to monitor the domestic online information environment for foreign disinformation affecting federal democratic institutions outside an election period. The proposal itself, however, is a private LinkedIn article by Dave McMahon, not a government strategy, program, appropriation, requirement or operational capability. 1 2
McMahon’s Sept. 8 article calls for a private-sector-led, federated cognitive-defence network; a National Centre for Cognitive Resilience reporting to the Privy Council Office; a threat-intelligence exchange; platform algorithm audits; and investment in deepfake detection and provenance tools. It also advocates early warning, public resilience and foreign-facing disruption. Those are proposals, not confirmed Canadian obligations or authorities. 1
Canada already has operational and educational pieces of the problem set. Global Affairs Canada’s Rapid Response Mechanism Canada, or RRM Canada, has monitored the digital environment for foreign information threats since 2018, conducts multilingual open-source research and analysis, and supports federal-election protection through the Security and Intelligence Threats to Elections Task Force. The Privy Council Office’s Protecting Democracy Unit also provides federal public servants with a guidebook and training resources, while government toolkits are available to community leaders and elected officials. 3 4
The distinction matters. The RRM’s stated work is foreign information manipulation and interference, or FIMI; it is not a public license to judge all domestic political speech or establish a universal domestic surveillance program. The Hogue Commission said a standing domestic monitoring function would require a legislative framework that respects privacy rights. It recommended considering an entity to monitor open-source domestic information for misinformation and disinformation that could affect democratic processes, structured to comply with law and able to exchange information with appropriate partners. 3 5
| Item | What the public record confirms | What remains a proposal, gap or test |
|---|---|---|
| Threat monitoring | RRM Canada monitors the digital environment for foreign threats detrimental to Canadian interests and has open-source monitoring expertise. 3 | A standing, specifically authorized domestic open-source function outside elections remains a policy and legislative question identified by the Commission. 5 |
| New institution | No announced National Centre for Cognitive Resilience or Canadian Cognitive Threat Intelligence Exchange was found in the cited government records. | McMahon proposes both structures and a federated private-sector network. 1 |
| Platform and AI tools | Canada offers disinformation guidance and says foreign actors can use social platforms and generative AI. 3 6 | Algorithm audits, deepfake-detection procurement and provenance systems are suggested capabilities; no outcomes record or deployment plan is supplied in the proposal. 1 |
| Economic harm | Government materials document discrete harms, including more than $300 million lost to investment fraud in 2022. 6 | McMahon’s C$50 billion to C$120 billion annual estimate has no reproducible methodology in the article; it should not be treated as a validated national cost ceiling. 1 |
A narrow public mandate is the workable part of the case
The strongest overlap between the proposal and the official record is narrow. Commissioner Marie-Josée Hogue recommended a legal framework for collecting and assessing open-source domestic intelligence with privacy protections, followed by consideration of a compliant government entity focused on misinformation and disinformation affecting democratic processes. She expressly said she was not recommending monitoring all social-media activity or private and semi-private communications of Canadians. 5
That leaves room for a defined pilot, but not for assuming a national network may simply begin “sensing” domestic audiences. Under the federal Privacy Act, a government institution may collect personal information only where it relates directly to its operating program or activity. It generally may use that information only for the purpose for which it was obtained or a consistent use, and may disclose it without consent only under the Act’s specified exceptions. 7 The Charter also guarantees freedom of thought, belief, opinion and expression, including press and other media, subject only to justified limits prescribed by law. 8
Those are not procedural details. They determine whether a program aimed at identifying coordinated foreign conduct stays focused on attributable foreign activity, public content and demonstrable threat indicators, or drifts toward mapping lawful Canadians’ beliefs, associations and political debate. A contractor’s role would not remove the need for a clear public mandate, data rules and accountable decision-making when it performs functions for the state.
Sweden’s Psychological Defence Agency offers a more limited comparator than the proposal suggests. Its published mission is to coordinate and support other actors, identify and analyze malign information influence directed at Sweden by antagonistic foreign powers, and strengthen resilience. The agency says it does not monitor or register what is expressed or by whom in Sweden’s information environment, and describes safeguarding freedom of expression and an open society as the purpose of psychological defence. 9 That is a useful boundary for comparison, not proof that a Canadian agency with broader authority would work.
Evidence supports concern, not a quantified return on investment
There is ample evidence that Canadians encounter questionable online information. Statistics Canada reported that 73% of Canadians in the 2022 Canadian Internet Use Survey had seen online content they suspected was false or inaccurate in the previous 12 months; 59% said they were very or extremely concerned about online misinformation in 2023. The agency cautioned that its surveys did not directly measure respondents’ ability to identify misinformation, which limits conclusions about real-world effects or the performance of countermeasures. 10
The government likewise reports concrete but bounded harms. It says Canadians lost more than $300 million to investment fraud in 2022, including victims of deceptive social-media advertising and cryptocurrency scams. That figure cannot be scaled into a national FIMI-loss total without a stated causal model, population, time period and counterfactual. 6
McMahon’s article labels its C$50 billion-to-C$120 billion annual estimate “conservative,” but cites a prior LinkedIn post rather than a transparent study. The article does not define what losses are included, explain how foreign manipulation is separated from domestic misinformation, fraud, poor information or market conditions, or show how double counting is prevented. It therefore is a claim, not a verified baseline or a funding ceiling. 1
Independent scholarship gives a reason to avoid treating the term “cognitive warfare” as a ready-made measurement system. A 2024 conceptual analysis found the NATO Allied Command Transformation concept broad and difficult to apply empirically because of its many attributes and limited real-world instances. A 2025 legal analysis of AI-supported countermeasures concluded that causality and attribution are difficult to establish and that state intervention must be based in law and strictly proportionate. 11 12
What a pilot would have to demonstrate
A defensible next step would be a time-limited, legislated pilot centered on public, open-source evidence of coordinated foreign information operations affecting defined democratic processes. It should not be authorized to label domestic criticism, disagreement, satire or unpopular speech as hostile activity. The pilot’s aim would be to improve detection, attribution and warning—not to “command” a national narrative.
Before operational use, Parliament and the responsible minister should publish the program’s statutory purpose, data categories, retention periods, disclosure paths, platform-contact rules, attribution threshold, correction process and independent review arrangements. A public vendor register and conflict-of-interest rules would be essential if commercial analytics firms provide tools or data. These are conditions for a test, not confirmed obligations in the present record. They follow from the Commission’s privacy-first recommendation and from the existing legal framework. 5 7 8
The pilot should be judged against pre-announced measures that can fail. Relevant tests include whether analysts correctly distinguish confirmed coordinated foreign activity from ordinary domestic conversation; whether warnings reach affected institutions in time; whether public attributions are later sustained by disclosed evidence; how many false positives are corrected; how much identifiable information is retained or shared; and whether independent privacy, legal and technical reviewers find noncompliance. A rise in alerts, content removals or media mentions would demonstrate activity, not protective effect.
Canada’s official response already embraces portions of this design: RRM Canada uses open-source analysis, discloses operations where appropriate and works with government, industry and international partners; the G7 framework emphasizes national strategies, governance, capacity, civil society and multilateral engagement while calling for expression safeguards and protection of marginalized groups. 3 13 The missing evidence is not that foreign manipulation exists. It is whether the much broader Canadian network McMahon proposes can improve resilience at an acceptable cost without weakening the privacy, expression and institutional trust it aims to defend.
References
1 Dave McMahon, “A National Cognitive Defence Strategy for Canada”
2 Land Offset, “Canada’s defense narrative challenge is bigger than messaging”
3 Global Affairs Canada, “Rapid Response Mechanism Canada”
4 Government of Canada, “Protect yourself and others from disinformation”
5 Public Inquiry Into Foreign Interference, Final Report, Volume 5
6 Government of Canada, “Online disinformation”
7 Justice Laws Website, Privacy Act, sections 4–8
8 Justice Laws Website, Canadian Charter of Rights and Freedoms, sections 1–2
9 Swedish Psychological Defence Agency, “Our mission”
10 Statistics Canada, “The spread of misinformation”
11 Christoph Deppe and Gary S. Schaal, “Cognitive warfare: a conceptual analysis of the NATO ACT cognitive warfare exploratory concept”
12 Henning Lahmann, Bart Custers and Benjamyn I. Scott, “The fundamental rights risks of countering cognitive warfare with artificial intelligence”
13 Global Affairs Canada, “The Framework to Counter Foreign State Information Manipulation”