Canada’s innovation ministry has released the results of the 30-day national consultation that will inform the country’s renewed AI strategy. However, the report’s high-level summary cites little of the data or methodology behind the compilation of its broad recommendations.
While the report states that respondents “strongly emphasized the need for Canada to attract, retain and develop top AI talent,” it isn’t clear who, or how many, held this belief.
The public AI consultations were held in October of 2025. They gathered input from AI minister Evan Solomon’s AI strategy task force, as well as a record number of submissions from the public.
Solomon planned the task force and public consultations with the intention of tabling a renewed AI strategy by the end of 2025. He later pushed the timeline for the release of the report to 2026, given the need to absorb more than 11,000 submissions.
The government’s summarized takeaways echo much of what Solomon and Prime Minister Mark Carney advocated for leading up to the consultations, including the need to build sovereign compute and data infrastructure, converting pilot projects to real deployment, and improving access to capital, procurement opportunities, and partnerships.
But the methodology for extracting these takeaways is unclear. The government says that it used multiple AI tools to interpret the data, analyzing responses with the Canadian enterprise survey tool SimpleSurvey, before using an in-house classification pipeline for analysis that incorporates large language models like Cohere’s Command A, OpenAI’s GPT-5 nano, Anthropic’s Claude Haiku, and Google’s Gemini Flash to read through the submissions and identify common themes.
ISED says it developed a “scalable, AI-enabled workflow,” called a classification pipeline, that used several large language models, including Canadian models, to clean survey responses and categorize them into a structured set of themes and subthemes. It says it used manual human review at several stages to ensure that “intents were meaningful and sensible and that the solution had at least a 90% success rate in categorizing responses into specific intents.” However, ISED does not clarify which prompts or workflows were used in that pipeline, or how it determined what constituted classification success.
RELATED: Canada’s AI ambitions with Minister Evan Solomon
ISED said this method sped up the process of tabulating more than 64,000 responses. However, the results appear as sweeping takeaways and do not cite any weighting of the data nor specific sources. For example, while the report states that respondents “strongly emphasized the need for Canada to attract, retain and develop top AI talent,” it isn’t clear who, or how many, held this belief. Similarly, it states that respondents hold concerns about “premature deployment and overhyped technologies like generative AI,” as well as “environmental harm, privacy risks and job displacement,” but it does not state how many respondents expressed these concerns. BetaKit is in the process of reviewing the full report to compare the public submissions to the AI-compiled takeaways and recommendations.
Concerns about consultation process
Surveys have shown that Canadians are generally divided on AI. Last year, Abacus Data found that 61 percent of Canadians believed AI posed a threat that could harm employment, personal privacy, and the stability of Canadian society. Fifty-two percent also said they do not trust the federal government to oversee AI in a way that protects the public.
The report’s overall considerations also noted that Canadians and Indigenous communities are “concerned that AI threatens control over creative works,” and that respondents called for stronger IP laws, opt-in consent, fair compensation, and protection for vulnerable creators.
“Respondents and stakeholders across industry, academia and government emphasized the need for a balanced approach—one that drives innovation, while safeguarding sovereignty, ethics and public trust, and adopts evidence-based decisions that prioritize public interest and democratic values,” the report states.
The findings are anchored around the eight pillars Innovation, Science, and Economic Development (ISED) Canada outlined going into the process: research and talent; industry and government adoption; commercialization; scaling and investment; safe AI and public trust; education and skills; infrastructure; and security.
Jaxson Khan,
“[The] diagnosis is often remarkably consistent, but the prescriptions are not.”
Aperture
ISED did provide demographic data of the respondents who chose to share it. That data indicated that more than one-third of people who disclosed their industry work in IT, technology, or cybersecurity, and that a plurality of respondents who disclosed their age group (30 percent) were between 35 and 44 years old. Other prominent groups of respondents included those from professional, scientific, and technical services (20 percent), arts, entertainment, and recreation (15 percent), and academia (13 percent).
There have been public concerns about the federal government’s consultation process, including an open letter that criticized the short timeline to take in feedback and the number of industry players on the AI task force. The open letter spawned its own competing public consultation, which is open until mid-March.
As for the AI strategy task force, ISED received 32 reports from 28 members across groups from tech to academia, as well as close to 300 supplementary policy papers or ad hoc submissions from business groups, governmental organizations, and NGOs.
Aperture AI CEO Jaxson Khan, a former policy advisor for ISED, said in a LinkedIn post that he looked through some of the submissions from the AI task force. He concluded that the “diagnosis is often remarkably consistent, but the prescriptions are not.” Khan added that this raised tensions between picking winning companies versus building broad, sustainable ecosystems, or moving fast versus taking time to build trust.
“These aren’t rhetorical questions but in some cases genuinely different theories of how Canada should position itself and what we should prioritize,” Khan said.
Have you read through Canada’s AI strategy submissions, or do you have thoughts of your own? Let us know what you think.
Feature image courtesy SAAS NORTH.