A federal advisory group is urging Canada’s security agencies to provide clear details about how they currently use and plan to use artificial intelligence systems and software. The Canadian Press
A federal advisory group is urging Canada's security agencies to be more transparent about their use of artificial intelligence (AI). The National Security Transparency Advisory Group has recently released a report recommending that these agencies provide detailed information on how they currently use and plan to use AI technology. This call for transparency aims to enhance public understanding and oversight of national security practices.
The advisory group, established in 2019 to promote accountability and public awareness of national security operations, stresses that while security agencies are already employing AI for tasks such as document translation and malware detection, more clarity is needed. AI is expected to play a growing role in analyzing vast amounts of data, identifying patterns, and understanding trends. The report highlights the importance of public knowledge about these technologies and their applications in border control, policing, and intelligence.
The report also suggests that the government should consider updating existing legislation to ensure proper oversight of AI usage by federal agencies. Currently, there is a requirement for federal entities to assess the impact of algorithms before they are used, and the proposed Artificial Intelligence and Data Act is intended to guide the responsible development and deployment of AI systems. However, this act does not currently apply to national security agencies, leading the advisory group to recommend extending its scope to include them.
The Canadian Security Intelligence Service (CSIS) and the Communications Security Establishment (CSE) have both responded to the report, emphasizing their commitment to transparency while acknowledging the challenges posed by their secretive work. The CSE, for instance, has used AI to enhance its capabilities in translating languages and detecting cyber threats. It has developed new AI models and tools, including a translation system that can handle over 100 languages. Despite its efforts to contribute to public understanding of AI, the CSE faces limitations in sharing detailed information due to national security concerns.
Similarly, CSIS has welcomed the advisory group's recommendations and is working on formalizing its approach to AI use. However, the agency notes that operational secrecy limits how much it can disclose publicly to protect its effectiveness and security.
The report also draws attention to past controversies, such as the RCMP's use of facial recognition technology from Clearview AI. This technology, which collects and analyzes large numbers of images, raised privacy concerns and legal issues. In response, the RCMP is developing a new Technology Onboarding Program to ensure that its use of technology complies with privacy laws. The advisory group urges the RCMP to be more open about this program to improve transparency.
Overall, the advisory group calls for a formal review of the government's transparency efforts, including public reporting on the progress and impacts of its commitments. Public Safety Canada has acknowledged receipt of the report and shared its recommendations with relevant officials and committees, but has yet to outline specific actions or timelines for implementing the suggestions.
