Artificial Intelligence (AI): Once the stuff of science fiction, AI is now the basis of science fact—especially when it comes to modern military technology. In this first CDR AI Report, we will look at the realities of AI and its potential role in the defence of Canada—and how this software advance is changing the very nature of military technology. We will do so by speaking with Canada’s CAF/DND Defence Team who are leading the charge, along with some of this country’s top AI, defence, and intelligence firms to get their insights. Here is our report.
BY JAMES CARELESS
Canada’s DND/CAF AI Strategy is being led by the Defence Team, a unified body combining personnel from the Canadian Armed Forces (CAF) and the Department of National Defence (DND). According to the Defence Team, the primary operational goal of AI is to speed up the CAF’s reaction cycle and reduce the cognitive load on decision-makers.
The five pillars of the AI Strategy encompass the lines of effort that DND must undertake to become AI-enabled by 2030. These pillars apply to the use of AI in the Canadian corporate as well as the military space.
According to the AI Strategy, the five pillars are: • Fielding and employing AI capabilities; • Change management; • Ethics, safety and trust; • Talent and training; • Partnerships.
“Artificial intelligence is a generational opportunity for both the Canadian Armed Forces and Canada as a whole,” said Ross Ermel, DND’s Assistant Deputy Minister (Digital Services) and Defence Chief Digital Officer. “For Defence, AI is a critical force multiplier that strengthens decision-making, readiness, and operational effectiveness; for the country, it is a cornerstone of economic growth, innovation, and digital sovereignty. Through the DND/CAF AI Strategy and the work of our DND/CAF AI Centre (DCAIC), we are building sovereign capabilities, investing in Canadian research and industry, and ensuring that all Defence uses of AI are governed responsibly, with strong oversight, clear accountability, and human judgment at the centre, reflecting Canadian values and the rule of law.”
To ensure that Canada’s AI adoption aligns with Canadian values, democratic accountability, and civilian control of the military, the DND/CAF AI Strategy emphasizes responsible AI adoption, with sovereign data governance as a core foundation of that responsibility.
To balance centralized AI governance with the need for experimentation and innovation at the unit or operational level, Canada’s Defence Team has adopted a federated approach to AI, where accountability for AI initiatives remains with the organizations closest to the mission, supported by enterprise-level governance. As well, the DCAIC is working with industry, academia, and partners to build sovereign AI capabilities for CAF warfighters.
According to the Defence Team, the demand for AI expertise exceeds supply across Canada and globally. Gaps are most evident in highly specialized technical roles, and in the operational and leadership capacity required to translate AI into a deployable capability.
In response, the Defence Team is making unprecedented investments through the Defence Investment Agency and other mechanisms to stimulate, scale, and sustain Canada’s AI ecosystem, working hand-in-hand with industry partners. These investments help Canadian firms attract, train, and retain AI talent, while accelerating the development of capabilities.
As well, the Defence Team is pursuing partnerships with industry and academia, to move the development and deployment process along faster. It is establishing an annual AI conference to bring together Canadian companies to showcase capabilities and build future partnerships, and to discuss Defence Team priorities directly with industry. The Defence Team is also supporting ISED innovation pathways to help Canadian startups mature, to create opportunities for companies of all sizes to become partners in Defence.
Worth noting: Canada’s Defence Team is already transitioning from experimentation to more sustained use in select areas where AI has demonstrated value, reliability, and trustworthiness. For example, the Canadian Army is now using AI to enhance virtual and augmented reality simulations to deliver more realistic and complex operational scenarios. However, the Defence Team notes that success is contingent on data hygiene, and that early integration is being prioritized in areas where data is already well-organized, such as Command and Control.
The takeaway: “Operationally speaking, AI enables the Defence Team to act faster, smarter, and with greater confidence, while reinforcing Canada’s broader industrial and innovation objectives,” said Major-General C.J. Zimmer, DND’s Chief Digital Transformation and Defence Chief Data Officer. “As these technologies are adopted, our priority remains clear: AI must support human decision-making, comply with Canadian and international law, and earn the trust of Canadians by being legal, ethical, and safe in every Defence application.”
That said, what keeps Canada’s Defence Team up at night is AI deployed without sufficient safeguards, where automated systems could increase the risk of miscalculation, unintended escalation, or loss of human control. This is why Canada’s approach to military AI is deliberately different. Its goal is building AI-enabled defence capabilities while preserving human accountability, legal compliance, and alignment with allied norms, even as the strategic environment becomes more contested and complex.
If the Defence Team defines the “why” and the governance of military AI, Canada’s defence industry is defining the “how” — and at what speed.
ANVIL INTELLIGENCE
Formerly known as IMRSV Data Labs, Anvil Intelligence has partnered with national defence organizations for over a decade through products like its ANVIL AI-enabled data management platform.
“The core problem we solve is managing the exponential growth in the volume and velocity of mission-critical data,” said Samuel Witherspoon, Anvil’s CEO. “Some of our initial work involved applying machine learning to various use cases for the federal government and that work evolved into the ANVIL platform, which is deployed with the CAF today.”
Asked to assess the DND/CAF AI Strategy, Witherspoon observed that, “The key signal which will drive investment… is how any AI strategy is executed through working with sovereign Canadian companies.” He defines these strictly as firms that are “Canadian-owned, own their IP, and reinvest their profits here.”
Regarding procurement challenges—and Ottawa’s attempts to accelerate deployment via the new Defence Investment Agency (DIA)—Witherspoon points to a proven model. “Anvil has shown that there is a successful path in Canada from pilot to fielded capability,” he told CDR. “Future development of successful AI products for the CAF will rely on similar close collaboration.”
Witherspoon sees this collaboration as central to the strategy’s success. “Industry and the CAF need to be collaborating on the hardest problems we know will require sovereign solutions,” he said. “There are already many public current and upcoming major capital projects stretching into the next decade which Anvil and other sovereign AI companies will need to contribute to.”
Conversely, “the overarching AI threat to Canada is that we don’t develop a strong, domestic technology sector,” he warned. “Without this, we won’t be able to generate the talent and capabilities which will form the foundation of our resilience.”
Witherspoon noted that the answer is already in the 2025 budget, adding: “The domestic defence ecosystem will only grow if the Canadian government becomes its own best customer and buys from Canadian companies who are developing critical capabilities for CAF-DND.”
COHERE
Cohere has positioned itself as a “security-first enterprise AI company,” with its primary defence relevance founded in a strict adherence to privacy and regulatory compliance.
“We build frontier AI models and agentic systems designed to solve real-world problems for enterprises and governments,” said Dave Ferris, Cohere’s VP of Americas and Global Public Sector. “Our state-of-the-art AI solutions enable organizations in the defence sector to enhance decision-making, operational efficiency, and mission effectiveness. In particular, all our products are designed to meet the highest security, privacy, and regulatory compliance standards—criteria that are non-negotiable in defence, making Cohere an ideal partner for government and industry leaders.”
In the defence sector, Cohere offers a comprehensive suite of AI solutions. These include enterprise-grade generative and search AI models, and the company’s ‘North’ flagship agentic AI platform to automate repetitive tasks.
“We are actively engaged with defence partners to explore how our AI technology can address critical operational challenges and strengthen Canada’s national defence,” Ferris told CDR. “As our technology is foundational, these solutions support both institutional and operational workflows, from the battlefield to back office. On the operational side, we focus on direct support to command and control, ensuring operators and analysts have immediate access to the information they need to understand their environments, make timely decisions, and take appropriate actions.”
According to Cohere, the most significant threats facing Canada and its allies concern compromised data sovereignty, adversarial AI manipulation, cyber threats to critical infrastructure, and autonomous weapon system proliferation.
“To address these challenges, it’s important that we accelerate adoption of AI across the defence sector and ensure Canada remains a global leader in responsible AI development,” said Ferris. “Our partnership with the Government of Canada is a key example of leveraging this transformative technology to advance national security.”
As to the federal government’s role? “Canada has a generational opportunity to become a world leader on AI development and deployment,” he replied. “Making an intentional choice to modernize the defence industry with responsible and secure AI has the potential to provide immense domestic economic and national security benefits. As Canada increases its investment in new capabilities, it is essential to consider and include Canadian AI systems and technologies.”
CONTEXTERE
Contextere describes itself as an industrial AI company. Its development traces back to a 2015 partnership with Lockheed Martin and the RCAF’s CC-130J program. More recently, Contextere has supported the RCAF’s CT-142 and CP-140 fleets.
“Our flagship product, Madison, is a domain-specific insight engine and Blue Collar Co-Pilot that turns siloed data from manuals, sensors, and enterprise systems into real-time, contextual guidance for technicians on the last tactical mile,” said CEO Gabe Batstone. To make this happen, Madison uses proprietary machine learning technologies to ingest and process the highly-technical, dynamic datasets common in military maintenance, field service, and operations.
“Madison’s algorithms are refined to extract nuanced meaning from dense technical documents, capturing specialized terminology, acronyms, and domain-specific language,” Batstone told CDR. “Its multilingual embeddings and vectorization techniques ensure that insights are both accurate and relevant, regardless of the user’s language or the complexity of the source material.”
Batstone describes Canada’s AI Strategy as an “important and credible signal,” but warns that to catalyze investment, it must address the present and future simultaneously.
“In the near term, government needs to back innovative AI programs with rapid funding, clear pathways to contract, and an acceptance that meaningful advantage comes from managed risk,” he told CDR. “When industry can see both immediate opportunities to build and deploy today, and a clearly defined future architecture to scale into, the signal shifts from intent to commitment.”
Batstone argues that current procurement models only partially align with modern AI cycles, often trapping companies in “pilot purgatory.” He calls for structural changes that prioritize deployable capabilities over broad academic collaborations. Meanwhile, the most realistic near-term opportunities for CAF AI deployment lie in capturing institutional knowledge from retiring experts and improving asset availability through predictive maintenance.
On the threat landscape, Batstone highlights the risk of over-reliance on autonomous tools without human oversight, as well as the danger of government staff using unsecured public AI tools.
Overall, Batstone’s advice to Ottawa is clear. “Accelerate the procurement process to expand the defence industrial base,” he said. “Concurrently, they need to reinforce ‘Made in Canada’ strategies for AI applications and systems to ensure that data assets and intellectual property remain under Canadian control.”
NEXTRIA
Nextria is focused on secure, on-premise generative AI for defence, intelligence, and other highly regulated environments. Since 2020, it has worked with the Department of National Defence (DND) and Defence Research and Development Canada (DRDC) on AI research programs, policy and standards chatbots, and machine-learning initiatives used by the CAF.
“Our core platform, CastleGuard AI, is currently used in defence and intelligence contexts for secure translation, transcription, and knowledge discovery across sensitive datasets,” said CEO Eric Kitchen. “A key component is Evia, our Canadian French/English expert language model built specifically for Government of Canada and defence terminology. We also deploy CastleGuard Insight, which enables secure ingestion, correlation, and analysis of large volumes of unstructured data in air-gapped and classified environments.”
According to Kitchen, the most realistic near-term opportunities for AI-enabled defence solutions are not the most exotic problems. Instead, “They are the everyday, repetitive tasks the CAF struggles with: document processing, translation, reporting, information retrieval, and basic decision support,” he told CDR. “Small, embedded teams of AI practitioners working directly with CAF personnel can deliver immediate value while building trust and understanding. Start with the basics, train users, and build confidence. If that foundation is not solid, more advanced AI efforts will fail.”
Kitchen views the CAF’s AI strategy as “positive and necessary,” but ultimately “evolutionary rather than transformational.” While it addresses AI usage issues such as ethics and governance, the AI Strategy “does not yet provide, at least publicly, a clear picture of which operational problems will be solved first, how quickly capabilities will be fielded, and how successful pilots will be scaled,” he observed. “Without that clarity, industry struggles to justify long-term investment, especially smaller companies.” He also highlights a strategic blind spot regarding threats, noting that Canada underestimates the risk of adversarial AI agents manipulating information narratives.
Meanwhile, Kitchen views Canada’s procurement systems as being optimized for risk avoidance—a stance misaligned with modern technology. “AI is not a one-time purchase. It requires continuous iteration, access to data, rapid updates, and tight feedback loops,” he told CDR. “We need to shift from diagnosing problems to delivering solutions: supporting innovators, accelerating execution and demonstrating that Canada can conceive, build and deploy world-class artificial intelligence domestically.”
Sapper Labs Group (SLG) is a veteran-owned, sovereign intelligence and cyber defence company that has supported national security missions for two decades.
“We deliver Intelligence and Infrastructure-as-a-Service, including managed attribution, multi-mission analytics, cyber operations capabilities, and security overwatch,” said Allen Dillon, CEO & Chairman. “Our Next-Gen agentic-AI powered multi-mission intelligence ecosystem accelerates the analytical process to manage risk and make informed decisions based on evidence-based facts. The ecosystem incorporates F3AEAD (Find, Fix, Finish, Exploit, Analyze, Disseminate) special forces targeting methodology, making it highly effective for joint information fusion and rapid targeting.”
SLG’s history includes a $11.3 million contract with the Innovation for Defence Excellence and Security (IDEaS) program (2020–2024), where the company used global commercial intelligence to attribute cyber adversaries. Looking ahead, SLG plans to launch a full multisource intelligence fusion platform in 2026, using agentic-AI to augment human analysis.
Dillon warns that AI is “reshaping how nations compete, coerce, and fight.” He identifies the “War on Truth” as perhaps the greatest threat, where adversaries use inauthentic AI to fuel disinformation and fracture democratic institutions. Other AI-enabled threats Dillon cited include enhanced hybrid warfare, cyber warfare conducted by AI-enabled systems, autonomous weapons systems, and total societal surveillance.
“Perhaps the most insidious threat is self-inflicted: Loss of strategic advantage without AI parity,” he told CDR. “Failing to appreciate and adopt AI at the pace of allies and adversaries. It requires sustained investment in data infrastructure, talent, experimentation, and allied interoperability—paired with strong governance to ensure responsible use.”
Critically, Dillon said that Canada’s current procurement model is simply too slow for the digital age. “Cyber-power now doubles roughly every ten months… Yet Canada’s cyber procurement cycles often take decades,” he observed. “The danger isn’t just efficacy; it is strategic exposure. Adversaries that iterate faster can exploit vulnerabilities long before democracies finish their requirements documents.”
His solution: Canada must establish a dedicated Director General for digital procurement. “We need a system that can procure nation-state intelligence and cyber systems for sovereign purposes without shining a light on our decisions for adversary intervention,” said Dillon. “If we are to realize the value of our AI investment… we must resolve the process issues. If not, we risk losing the IP and talent to the U.S.”
ZIGHRA
Zighra Founder and CEO Deepak Dutt describes his company as a “Canadian defence technology builder focused on developing operational, explainable AI for contested environments.” Zighra’s initial work focused on behavioural AI for identity and system assurance, including projects with the Royal Canadian Navy.
“Building on that foundation, our work expanded into explainable AI for situational awareness and cyber resilience, including work with the Canadian Coast Guard supporting decision-making across complex maritime and coastal operations,” Dutt told CDR. “More recently, our efforts have shifted toward AI-enabled sensing and electromagnetic awareness, driven by increasing spectrum contestation, navigation resilience concerns, and the rise of autonomous threats.”
Looking ahead, Zighra intends to bid on projects aligned with pressing CAF problem sets, including Counter-UAS sensing and decision support, electromagnetic spectrum awareness and protection, GNSS resilience in degraded environments, and AI-enabled fusion that supports timely, actionable situational awareness. “These are areas where Canada faces real operational pressure and where domestic, sovereign AI capability can provide durable advantage,” said Dutt.
Regarding the CAF’s AI strategy and its impact on industry, Dutt described its content as being “directional rather than decisive.” While the AI Strategy communicates the federal government’s intent and ambition, “on its own it is not yet sufficient to drive sustained private-sector investment, particularly for companies building defence-specific AI capabilities that require long development cycles and deep integration with military systems,” he said. “From our perspective as a Canadian defence technology builder that has delivered and tested AI capabilities in defence contexts, the most significant barrier is not technical readiness but transition friction. The distance between a successful demonstration and a fielded, sustained capability remains too great, and that uncertainty directly affects how much industry is willing to invest ahead of demand.”
To win industry support and to attract top-flight talent to the Canadian AI sector, “the strongest signal the government can send to talent is confidence in adoption,” said Dutt. “Engineers are far more willing to work on defence systems when they believe their work will move beyond pilots and be fielded, sustained, and improved in real operations. Clear pathways from testing to deployment matter as much for talent attraction as they do for procurement efficiency. If Canada aligns mission clarity, adoption pathways, and operational access, it can compete effectively for AI talent.”
James Careless is CDR’s Ottawa Bureau Chief
