Ottawa Launches Research Institute on AI Security

Innovation, Science, and Industry Minister Francois-Philippe Champagne (R), and founder and scientific director of MILA-Quebec Artificial Intelligence Institute Yoshua Bengio answer questions from the media following their announcement to launch the Canadian Artificial Intelligence Safety Institute in Montreal, on Nov. 12, 2024. The Canadian Press/Christinne Muschi

The federal government has opened a research institute focused on studying the security risks posed by artificial intelligence.

The Canadian Artificial Intelligence Safety Institute (CAISI) was created as part of the federal government’s broader strategy to address security risks associated with AI, including misuse in disinformation, cybersecurity breaches, and election interference, according to a Nov. 12 press release.

The creation of CAISI is part of the federal government’s $2.4 billion strategy for AI development as safety, as outlined in Budget 2024. The government says the country had more than 140,000 AI professionals in 2022–23, which is a 29 percent increase from the previous year.

CAISI, which will be overseen by the Innovation Department, has an initial budget of $50 million over five years.

AI-Driven Threats

The government has warned that threat actors, many backed by states such as China, Russia, and Iran, are increasingly using AI to target Canadians and interfere with the country’s democratic institutions. This includes a 2023 “spamouflage” campaign connected to China, which targeted dozens of MPs, including Prime Minister Justin Trudeau and Opposition Leader Pierre Poilievre.

“Spamouflage” is a tactic that uses networks of hijacked or fake social media accounts to spread propaganda. Global Affairs Canada (GAC), which detected the campaign in August 2023, confirmed the use of AI, including “deepfake” videos digitally altered by the technology, to target Canadian officials.

The incident was raised when executives from Google and the social media platform X testified before the House of Commons Ethics Committee on Oct. 24 about the growing threat of disinformation targeting parliamentarians. They identified China as the top source of foreign interference and disinformation campaigns aimed at Canada.

Lindsay Hundley, global threat intelligence lead for Meta, the parent company of Facebook, also highlighted that Chinese campaigns operate on a global scale and have targeted Canadian audiences as part of their efforts.

“China-origin operations have evolved significantly in recent years to target broader, more global audiences, including in languages other than Chinese,” she said while also testifying before the ethics committee. “These operations have continued to diversify their tactics, including targeting critics of the Chinese government, attempting to co-opt authentic individuals, and using AI-generated news readers in an attempt to make fictitious news outlets look more legitimate.”