Cabinet and Senate Pursuing AI on Parallel Paths
In his capacity as chair of the Canadian Internet Society, Brent Arnold has both contributed to the federal government’s public consultation on AI and testified before the Standing Senate Committee on Transport and Communications as it conducts its own study, giving him an inside view of concurrent efforts to understand the opportunities and challenges around AI in the interests of developing a national strategy. Here he shares his impressions of two very different processes.
The regulation of generative AI has been a contentious issue in Canada since before OpenAI released ChatGPT on an unprepared public, and despite the Trudeau government’s attempt to catch up to jurisdictions like Europe with the abortive Artificial Intelligence and Data Act, we still have no purpose-built federal law to show for it. Under the new Carney government, Minister Evan Solomon has pursued a public consultation on how Canada should approach artificial intelligence, while the Standing Senate Committee on Transport and Communications has been conducting its own study. As Chair of The Canadian Internet Society, I oversaw the submission of a written brief to the Minister’s consultation, and testified before the Senate committee. Thus, we at the Society have had a window into both projects.
The two processes have unfolded in very different ways. Minister Solomon conducted a “30-day sprint” consisting of a few invitation-only roundtables and the collection of written submissions on short notice. Written submissions were reviewed by AI chatbots and compiled into a “what we heard” report purporting to accurately reflect the opinions received. This process has proven controversial, in part, because people and organizations who respond to government calls for submissions expect them to be read by the government, not AI, and in part, because Professor Michael Geist replicated the process and discovered that the “what we heard” report appears to downplay criticisms of the way the federal government has handled the AI brief, as well as hiding a lack of consensus over trust and safety. It does not help that this process follows the decision not to resurrect the AI portion of Bill C-27, which itself was the product of minimal consultation and widespread criticism for its lack of specificity on the most basic of points (like, to what sort of AI applications the law would actually apply).
All in all, the government’s approach appears to have resulted in a hurried process to produce a report that has heightened distrust of the government’s intentions (and, possibly, its competence).
By contrast, the Senate committee’s study has set a relatively focused mandate with a few guiding questions (like how to address the intellectual property rights of creators whose works are used in training sets), and has given itself till the end of 2027 to report its findings to the Senate. Hearings started back in 2025 and are ongoing. So far, the committee has heard from law professors, AI industry analysts, thinktanks, and government agencies tasked with signals intelligence and cyber security. It has also heard from AI doomers encouraging an outright ban on generative AI in Canada, and experts on online harms raising concerns about the dangers of AI for children and vulnerable adults. The tone of the hearing I attended was very much more focused on the “challenges” presented by AI than the “opportunities” (to borrow again from the study mandate).
A few consistent themes arose that will, I think, continue to weave throughout the Senate committee’s deliberations:
- Generative AI isn’t just another new technology. It has the potential to be as transformative and disruptive as the internet, if not more so.
- Integrating AI into Canada’s economy, culture, and society at large, and doing so safely, is going to both a generational and a whole-of-society project.
- There are inherent tensions between this technology and the privacy and property rights, and even the safety and wellbeing, of individuals, and our laws are going to need to be updated and supplemented to grapple with this.
It appears, to me at least, that there is also a tension on Parliament Hill about how to address generative AI, as the government focuses on economic opportunity and competition, while the Senate appears focused on understanding and minimizing harms. The differences in priorities and pace will make for an interesting legislative process.
