Canada’s Spy Watchdog Reviews AI Use in National Security

The National Security and Intelligence Review Agency (NSIRA) in Canada is undertaking a comprehensive examination of the use and governance of artificial intelligence (AI) within national security operations. This initiative aims to assess how security agencies define, implement, and oversee AI technologies in their activities.

In a letter addressed to key federal ministers and security organizations, NSIRA Chair Marie Deschamps emphasized that the findings from this review will provide critical insights into the application of emerging AI tools. This will not only guide future evaluations but also identify any “potential gaps or risks” that may require attention. The review encompasses various tasks that Canadian security agencies have been employing AI for, including the translation of documents and the detection of malware threats.

The review agency possesses the statutory authority to access all information held by departments and agencies under scrutiny, including classified materials, with the exception of cabinet confidences. The letter, which has been made available on the agency’s website, outlines that requests for information may involve documents, explanations, briefings, interviews, surveys, and access to systems. Deschamps indicated that the review could also consist of independent inspections of certain technical systems.

Several cabinet members received this letter, including Prime Minister Mark Carney, Artificial Intelligence and Digital Innovation Minister Evan Solomon, Public Safety Minister Gary Anandasangaree, Defence Minister David McGuinty, Foreign Affairs Minister Anita Anand, and Industry Minister Mélanie Joly. Additionally, it was sent to leaders of major security entities such as the Canadian Security Intelligence Service (CSIS), the Royal Canadian Mounted Police (RCMP), and the Communications Security Establishment (CSE).

The RCMP expressed its support for the review, highlighting the importance of independent examinations of national security and intelligence operations. In a media statement, the RCMP asserted that “establishing transparent and accountable external review processes is critical to maintaining public confidence and trust.”

In 2024, the National Security Transparency Advisory Group released a report recommending that Canada’s security agencies provide detailed descriptions of their current and future uses of AI systems. The group underscored the anticipated increased reliance on AI technology for analyzing large volumes of data, recognizing patterns, and interpreting behavior.

While CSIS and CSE acknowledged the need for transparency regarding AI, they noted limitations on what could be publicly disclosed due to their security mandates. The federal government’s principles surrounding AI usage advocate for openness regarding how, why, and when AI is employed, while also assessing and managing any potential risks to legal rights and democratic norms early in the process.

In its most recent annual report, CSIS indicated that it was implementing AI pilot programs consistently with the federal government’s guiding principles. The RCMP has outlined several factors essential for ensuring that AI is utilized legally, ethically, and responsibly. These include careful system design to prevent bias and discrimination, privacy considerations during information analysis, transparency in AI decision-making, and accountability measures to guarantee proper functioning.

The CSE has articulated its commitment to developing new capabilities to address critical challenges through innovative AI and machine learning technologies. The agency’s strategy suggests that when deployed effectively, AI can enhance its data analysis capabilities, leading to improved decision-making speed and quality.

Caroline Xavier, Chief of CSE, stated in the strategy message, “We will always be thoughtful and rule-bound in our adoption of AI, keeping responsibility and accountability at the core of how we will achieve our goals.” She further emphasized the importance of rigorous testing and evaluation, ensuring that trained experts remain involved in the process.

This review marks a significant step in understanding the intersection of AI technology and national security in Canada. As security agencies increasingly integrate AI into their operations, the findings may shape future governance and oversight frameworks for the responsible use of these powerful tools.