The Canadian government unveiled its ambitious Canadian Artificial Intelligence Safety Institute (CAISI) yesterday. Minister of Innovation, Science and Industry François-Philippe Champagne announced the initiative at Mila, Quebec's Institute of Artificial Intelligence, highlighting an initial $50 million five-year budget to address growing concerns about AI risks, including disinformation, cybersecurity threats, and election interference.
"The Canadian Institute for Artificial Intelligence Security will propel Canada to the forefront of global efforts to use AI responsibly, and will be a key player in building public trust in these technologies. In a world that's evolving quickly and full of unknowns, Canadians can be confident that we will always take the necessary steps to ensure the AI they use is safe, responsible, and trustworthy," Minister Champagne stated during the launch event.
The institute will operate under two primary research streams:
- A CIFAR-managed programme funding applied and investigator-led research by Canadian and international experts
- Government-directed projects led by the National Research Council of Canada focusing on national priorities such as cybersecurity
CAISI will leverage expertise from Canada's existing AI infrastructure, including the National Research Council, CIFAR, and the country's three national AI institutes: Amii in Edmonton, Mila in Montréal, and the Vector Institute in Toronto.
The initiative comes at the right time for Canada's AI sector, which has shown remarkable growth across multiple metrics:
- Over 140,000 active AI professionals in 2022-23, up 29% from the previous year
- 67% growth in women in AI participation (highest globally) in 2022-23
- 57% increase in AI patents filed by Canadian inventors
- $8.6 billion in venture capital attracted to the AI sector in 2022
The institute will also participate in the International Network of AI Safety Institutes, with its first meeting scheduled for later this month in San Francisco, furthering Canada's commitment to the Bletchley Declaration on international AI safety collaboration.