Canada's long-awaited national artificial intelligence strategy, unveiled on June 4, promises to accelerate the adoption of AI systems across the federal government through a new Office of Digital Transformation. However, the strategy does not set a specific launch date or provide detailed operational plans for the office.
AI for All Strategy
The "AI for All" strategy, delayed for several months, outlines six pillars aimed at harnessing AI's economic potential while mitigating risks. Goals include creating thousands of AI-related jobs, helping small and medium businesses access public computing infrastructure, and using AI to reduce administrative burdens on public servants while keeping humans in decision-making roles.
Prime Minister Mark Carney announced the strategy at a Toronto news conference, stating, "Canadians need to be able to develop the trust that AI will be governed in ways that reflect our values. To trust that they will share in its benefits."
Office of Digital Transformation and Innovation Fellows
The strategy recommits to launching the Office of Digital Transformation, first mentioned in the federal budget over seven months ago, but provides no timeline or details on its mandate, staffing, or scope. It also promises an Innovation Fellows Program to recruit talent capable of building internal capacity for AI deployment across government.
"Canada will lead by example in responsible AI use within the public sector," the strategy states. "AI will be used to support public servants by reducing administrative burden and enabling better service delivery, all while maintaining a 'human-in-the-loop' approach to ensure strong oversight."
Union Concerns
The government's AI push has drawn criticism from unions. The AI-powered translation tool GCTranslate has raised concerns about linguistic rights and loss of nuance. Nathan Prier, president of the Canadian Association of Professional Employees, previously called AI a "Trojan horse" for public service cuts, warning that replacing rather than supplementing workers could lead to arbitrary reductions.
