OTTAWA — The national spy watchdog is investigating the use and governance of artificial intelligence by Canada's security agencies.
The National Security Intelligence Review Agency sent letters to key federal ministers and organizations informing them of the study.
Here's a look at how some federal agencies are leveraging AI tools to support national security, intelligence, and risk assessment operations.
Canadian Security Intelligence Service
CSIS works to protect Canada from espionage, terrorism and foreign interference. The spy service plans to launch a pilot project early this year to evaluate AI tools used to transcribe audio content, translate materials and examine documents to aid analysis, said spokesman Eric Balsam.
CSIS will also test chatbot-style tools to help create, edit, and summarize documents.
The output of AI systems is always subject to human review to ensure accuracy and appropriateness, Balsam said.
AI technologies being considered by CSIS will undergo algorithmic impact assessments to assess risk and manage potential adverse effects. Balsam said the spy agency is also seeking guidance from the Justice Department on legal considerations that would impact proposed uses of AI.
Royal Canadian Mounted Police
A 2024 RCMP report on operational technology said the force employed facial matching, a feature built into certain software applications used to process, classify and analyze large amounts of images and video.
“The RCMP will only use this technology for the purpose of processing evidence lawfully obtained during the course of an investigation,” the report states.
While AI offers the potential for efficient and innovative data processing and analysis in the context of policing and national security, it also presents certain challenges, said RCMP spokesperson Robin Percival.
These include ensuring the legal and ethical collection and use of personal information, preventing potential misuse or inaccurate use of information, and preventing access to data by unauthorized persons, he said.
The RCMP has established an AI policy and solutions team to develop a framework for the responsible deployment of artificial intelligence tools, Percival said. This includes creating a working group and drafting policies to allow RCMP employees to further benefit from new technology, she added.
Establish communication security
CSE, Canada's cyber espionage agency, uses artificial intelligence to protect federal government and critical infrastructure systems from cyber threats.
According to CSE's artificial intelligence strategy, this involves detecting patterns in vast amounts of data. “Machine learning enables the detection of a variety of threats and helps keep these systems and networks secure,” the strategy states.
The agency, which has long been at the forefront of technological innovation, also uses AI to classify malware.
“Machine learning-powered malware classification is critical to CSE as it enables the detection and analysis of advanced cyber threats that are not captured by standard antivirus solutions,” the strategy states.
Federal government networks are often targeted by attacks by foreign adversaries using custom malware, the document added, requiring more advanced detection capabilities than off-the-shelf antivirus software can provide.
According to the strategy, AI capabilities will become more important in CSE's efforts to protect Canadians in the coming years. “In particular, when deployed safely, reliably, and effectively, these capabilities will improve CSE's ability to analyze large amounts of data more quickly and accurately.”
Canadian international situation
The document search and analysis tool, called Document Cracker, was developed to help Canada's international affairs staff quickly find the information they need, according to a federal register of government-wide AI projects published last year.
“Document Cracker allows users to search large volumes of documents using an advanced indexing platform and automatically monitor new documents for mentions of specific topics, emerging trends, and key people, locations, and organizations,” the registry states. “These capabilities allow authorities to quickly identify pressing issues, develop preferred positions, and track the evolving positions of other countries.”
Canadian Immigration, Refugees and Citizenship
Federal passport programs use facial recognition technology to support identity verification, help detect fraud and prevent passports and travel documents from being issued to ineligible applicants, according to the Federal AI Registry.
Transport Canada
According to the AI Registry, the Risk Assessment and Conflict Tools Project is a data-driven initiative led by the department's Conflict Area Intelligence Directorate to improve Canada's ability to monitor and assess threats to passenger aircraft.
The tool was developed to streamline manual and resource-intensive processes such as open-source media monitoring, data analysis, and risk assessment through the use of automation and AI technology, the registry said.
This report by The Canadian Press was first published Jan. 7, 2026.
Jim Bronskill, Canadian Press
