Researcher gains access to governmental records on deliberations over Canada's failed ‘Artificial Intelligence and Data Act’
by Dr. Piyush Mathur
In a development useful for transparency and AI policy analysis, Blair Attard-Frost, a doctoral candidate and lecturer at the Faculty of Information, University of Toronto, has obtained 328 pages of previously undisclosed government records detailing the drafting process of Canada’s failed Artificial Intelligence and Data Act (AIDA). Attard-Frost announced the release on February 20, 2025, via a post on LinkedIn, sharing the full document package along with a spreadsheet breakdown highlighting key findings.
The Artificial Intelligence and Data Act—once a cornerstone of Canada’s push to regulate artificial intelligence—failed to secure parliamentary approval, leaving the country without a dedicated legislative framework to govern AI systems. Notably, AIDA was introduced in June 2022 as part of Bill C-27, the Digital Charter Implementation Act, 2022. However, on January 6, 2025, the Canadian Parliament was prorogued, resulting in all bills in progress, including Bill C-27, dying on the order paper.
This is a screenshot of Blair Attard-Frost’s LinkedIn post regarding her recent acquisition of governmental records relating to the drafting of AIDA.
According to Attard-Frost, these newly accessible records document internal deliberations, exposing ‘public engagement failures & process failures’ that contributed to AIDA’s collapse. ‘Though AIDA is now gone, these records have historical value’, she noted in her LinkedIn post, stressing that people should learn from these failures ‘to improve future AI legislation.’
The records were released following an Access to Information and Privacy (ATIA) request—a legal mechanism enabling public access to federal government documents. While the contents of the records remain to be fully analyzed, their publication will provide an insight into how Canadian policymakers grappled with regulating emerging AI technologies amid mounting public scrutiny and industry pressure.
The bill had critics and doubters from the beginning
AIDA faced significant challenges from its inception. Stakeholders did not think that sufficient public consultation had gone into drafting the bill; they also believed that it was overly focussed on promoting the Artificial Intelligence (AI) industry rather than protecting the rights and freedoms of Canadians.
There was also a concern that the bill granted excessive regulatory power to the Minister of Innovation, Science, and Industry—undermining parliamentary sovereignty. The proposed structure allowed the minister to establish and enforce regulations, including imposing shutdown orders and administrative fines, without clear rights of appeal.
Contrariwise, there were also significant concerns that the Artificial Intelligence and Data Act (AIDA) excluded AI systems used by government agencies from its scope. This exclusion raised substantial human rights concerns, especially given instances where government bodies, such as the Royal Canadian Mounted Police, have previously engaged in controversial AI practices, like the unlawful use of facial recognition technology. Critics argued that by exempting government institutions, AIDA failed to address the potential risks associated with AI applications in law enforcement and public safety. This omission was seen as a significant gap in the legislation, undermining its effectiveness in safeguarding individuals' rights against AI-related harms.
The legislation's ambiguity further compounded its challenges. Key concepts, such as ‘high-impact AI systems’, were thought not to have been clearly defined, leading to confusion about the scope and application of the regulations. Stakeholders claimed to struggle to understand their obligations and the prospects for the legislation’s effective enforcement.
Perhaps the overriding challenge that the bill had faced was that its developmental process had been viewed to be rushed, lacking in transparency and inclusivity. The absence of a broad-based public deliberation process meant that essential questions about the nature of AI, its associated challenges, and appropriate governance mechanisms were not adequately addressed. This oversight risked creating a regulatory regime misaligned with Canadian values and potentially damaging to the country's international reputation.
A window into future AI policy
Almost all of the foregoing concerns are reflected, in small doses, in the ‘Inventory of Records from ATIA Request - ISED AI Legislation’, rather carefully prepared by Attard-Frost; her inventory (in pdf here) also indicates exactly how many pages, from which parts of the 328-page dossier, have been held back by the government. The release of these records will help stakeholders reshape future debates on AI regulation in Canada. As Canada continues to navigate the ethical and regulatory challenges posed by artificial intelligence, these records may serve as a cautionary tale—and a roadmap—for future legislative efforts (not only in Canada but also in other countries sensitive enough to learn from Canada).
As for Attard-Frost, she has underscored the uncertainty of future AI legislation in Canada in an article published last month on the website of Montreal Ethics Institute. However, she also ended that article with a somewhat passionate pitch for a proactive, citizen-led effort at the local level to articulate AI regulation in Canada—instead of waiting for the next government (likely to be led by the Conservatives) to do something about it:
In the absence of clear and effective national AI regulation, we can organize with our co-workers and our communities to create smaller-scale policies, guidelines, and best practices for how AI should be built and used in the places that we live in and work. As Canada’s political upheaval continues and new regulatory norms for AI emerge, these bottom-up approaches to AI regulation will play an important role.
That is not all! Attard-Frost tops up her pitch with links to other publications (including one of her own) that she believes will help Canadians accomplish precisely that: Take the charge of AI regulation via the route of local needs and priorities in relation to AI.
Dr. Piyush Mathur is the author of Technological Forms and Ecological Communication: A Theoretical Heuristic (Lexington Books, 2017); for several years, he was an adjunct scholar with McMaster University’s Institute on Globalization and the Human Condition.
Background material
Attard-Frost, Blair (January 17, 2025) ‘The Death of Canada’s Artificial Intelligence and Data Act: What Happened, and What’s Next for AI Regulation in Canada?’ Montreal Ethics Institute (URL: https://montrealethics.ai/the-death-of-canadas-artificial-intelligence-and-data-act-what-happened-and-whats-next-for-ai-regulation-in-canada/)
Clement, Andrew (November 6, 2023) ‘AIDA’s “Consultation Theatre” Highlights Flaws in a So-called Agile Approach to AI Governance’ Centre for International Governance Innovation (URL: https://www.cigionline.org/articles/aidas-consultation-theatre-highlights-flaws-in-a-so-called-agile-approach-to-ai-governance)
Gruske, Carolyn (October 16, 2023) ‘Critics say Artificial Intelligence and Data Act needs to focus more on rights, not just business’ Canadian Lawyer (URL: https://www.canadianlawyermag.com/practice-areas/privacy-and-data/critics-say-artificial-intelligence-and-data-act-needs-to-focus-more-on-rights-not-just-business/380552)
Macdonald-Laurier Institute (MLI) (June 25, 2024) ‘Pulling the plug on federal AI Laws: The problem with the proposed amendments to AIDA’ (URL: https://macdonaldlaurier.ca/pulling-the-plug-on-federal-ai-laws-the-problem-with-the-proposed-amendments-to-aida/)
Sonja, Soloman & Christelle Tessono (December 6, 2022) ‘How To Fix Canada’s Proposed Artificial Intelligence Act’ Tech Policy Press (URL: https://www.techpolicy.press/how-to-fix-canadas-proposed-artificial-intelligence-act/)