Executive summary

Canadian government officials aim to deploy artificial intelligence (AI) responsibly. This report examines Canada’s consultation process on AI governance, comparing it with similar initiatives in Australia, Colombia, and the United States. The authors argue that democratic nations must find meaningful ways to engage citizens regarding the broad societal challenges presented by AI. While Canadian officials are working to improve public input processes, the report suggests democracies can do more to inform, involve, and collaborate with their citizens. Key recommendations include building a base of common knowledge about AI, recognising public participation as a marketing challenge, establishing an “always-on” feedback portal, and designating an ombudsperson within government departments to resolve citizen complaints related to AI policies.

Overview

On 26 September 2025, Evan Solomon, Canada’s minister of artificial intelligence and digital innovation, announced a 30-day “sprint” to seek public input on a national AI strategy. Key issues included attracting research talent, encouraging adoption across industry and government, scaling Canadian AI champions, building safe systems, and strengthening public trust. While the government invited a wide range of stakeholders, including workers, creators, and community voices, many groups criticised the abbreviated process. In October 2025, 41 Canadian organisations and 123 citizens sent a public letter arguing the process was unfair and “pre-cooked,” refusing to validate an outcome they believed was decided behind closed doors. They recommended extending the deadline to six months, reconstituting the task force, and rewriting the survey to remove bias.

Canada’s approach to participatory governance

Canada is viewed as a global leader in open government, yet it faces low levels of actual participation. An OECD 2024 Trust Survey found that only 40% of Canadians feel the political system allows them to have a say in government actions, and a study by the Privy Council Office reported that only 12% of Canadians believe people like them have a significant influence on politics. In 2024, the government adopted a Trust and Transparency Strategy to meet citizen expectations for accountable and participatory government. As part of this effort, the Treasury Board worked with the Morris J. Wosk Centre for Dialogue to create the Maturity Model for Public Participation, a framework to help agencies assess and strengthen their capacity for effective engagement. However, as of early 2026, there is no evidence this model has been operationalised, with pilot programmes slated to begin later in the year.

Methodology

The authors analysed five AI-related consultations held by Canadian officials between 2023 and 2025. The analysis focused on the process of obtaining and incorporating comments rather than the content of the comments themselves. Researchers addressed five key questions regarding participant demographics, materials provided for informed advice, outreach efforts, the timing of engagement, and evidence of how feedback was utilised. The report uses the International Association for Political Participation (IAP2) spectrum to characterise the level of participation, ranging from “Inform” to “Empower.” The findings and background data were compared against similar studies conducted in Australia, Colombia, and the United States to identify systemic challenges in democratic AI governance.

Summary of findings from the consultations

The report highlights findings from two primary consultations: one on copyright in the age of generative AI (2023–2024) and another on AI and competition (2024). The copyright consultation received 103 public submissions from organisations and approximately 1,000 questionnaire responses. Business and trade associations made up 46% of the 103 public submissions, while individuals accounted for only 13%. The competition consultation received 28 submissions, with 25 made public; of these, firms and business associations represented 28%, and individuals only 8%. For the broader national AI strategy, the government received 11,384 total responses, but 8,221 were “in progress” or unsubmitted. Of the completed submissions, 82% were from individuals, yet the sectoral breakdown showed that industry-linked groups and academics remained the most influential voices. Marketing for these consultations was limited to government websites and social media posts that received minimal interaction.

Conclusions and recommendations

The report concludes that while Canadian officials worked hard to be transparent, they have yet to achieve meaningful participation. Applying the IAP2 spectrum, both the copyright and competition consultations reached the “Inform” and “Consult” levels but failed to “Involve,” “Collaborate,” or “Empower” citizens. The authors offer several specific recommendations for democracies. First, governments should work to build a common knowledge base by providing accessible reading materials, such as Brian Christian’s The Alignment Problem or Ethan Mollick’s Co-Intelligence, to support informed debate. Second, they must view participation as a marketing problem, using diverse platforms and community organisations to reach a broader audience. Third, an “always-on” portal should be established at departments like ISED to allow constant feedback. Finally, an ombudsperson should be designated at every relevant agency to investigate complaints and recommend procedural changes, ensuring that AI governance is democratically determined rather than industry-led.