The government quietly released a “what we heard” report this week discussing the response to its 30-day sprint AI consultation from last October. Described as the “largest public consultation in the history of ISED”, the report relies heavily on AI for its analysis as the government notes that it used “Cohere Command A, OpenAI GPT-5 nano, Anthropic Claude Haiku and Google Gemini Flash to read through the submissions and identify common themes.” Given that it received 64,600 responses to 26 questions, it says AI enabled it to shrink a process that would typically take many months into a matter of weeks.
In addition to the public consultation survey, AI Minister Evan Solomon formed a 28 person expert committee that provided the government with 32 different papers and reports. Those documents were similarly subject to AI analysis with the “what we heard” report devoting several pages to the expert analysis and recommendations. Yet unlike the public survey responses, the government has posted all the experts’ reports, which allows the public to see the actual advice alongside the government’s summary of it.
Since the government used AI to summarize the expert reports, I thought I would do the same. I uploaded all 32 documents to both Chat GPT and Perplexity AI and asked for summaries of the major themes and areas of disagreement (the Chat GPT summary I generated is available here). While there are obviously many overlaps since these are summaries of the same documents, there are some notable differences that suggest the government had not provided the public with the full picture. Indeed, the direct advice from the experts that identifies policy choices and their implications is consistently softened into “government-speak” with balancing discussion that creates an illusion of consensus that isn’t really there.
For example, the expert reports consistently argue that Canada’s AI challenge is not about research excellence or talent creation, but rather execution. Whether commercialization of AI, AI adoption, infrastructure needs, or scaling to globally competitive operations, the reports repeatedly emphasize Canada’s failure to move beyond world-leading research. The government summary isn’t nearly as frank. Instead, it presents each policy pillar as if it were a parallel priority with balanced objectives. The result is a policy approach in which everything matters and few trade-offs are acknowledged. The experts are trying to sound the alarm on the risks to Canada if it fails to act but that isn’t the message the government seemingly wants to communicate.
The same is true for speed, which the expert reports frame as a strategic variable in which countries that move faster lead, while those that hesitate are left to regulate what others have built. The reports suggest government bears part of the blame here as some point to slow procurement, delayed funding decisions, and regulatory approval process barriers. Once again, you won’t find that in the government summary, where there is no acknowledgement that Canada’s slow pace may already be undermining its competitiveness, sovereignty, and capacity to shape global AI norms.
Access to capital and government procurement also get different treatments in the government reports as opposed to the actual expert reports. The government summary refers indirectly to the access to capital challenges without digging into the political choices. The expert reports are not nearly as circumspect as some believe that Canada’s inability to scale AI firms is tied to the absence of access to domestic capital, which is presented as a structural constraint, not a marginal concern. Similarly, some expert reports fear the reliance on grants has become counterproductive by shielding firms from market discipline while failing to generate customers, revenue, or scale. Government procurement is viewed as a far better industrial policy lever. In other words, compete for business rather than handouts. The government summary simply doesn’t engage with the issue.
Perhaps the most important divergence comes from the issue of trust and safety. This was a major concern from the public responses and the government is likely headed toward making AI governance, audits, transparency, and risk-based regulation key elements of its AI strategy. Yet there is far less consensus in the expert reports. Just about everyone agrees that trust is essential for AI adoption, but the implementation of regulation draws different views. Some want to move quickly, while others warn that overly broad regulation will slow deployment, disadvantage domestic firms, and regulate technologies Canada does not control. Those disagreements largely disappear in the government’s summary, where trust is presented as a settled consensus objective, rather than a contested policy domain with real trade-offs.
There are many other examples, including differing views on digital sovereignty, regulatory sequencing (what should be prioritized first), and inclusion. Put 28 experts on a panel and there is obviously going to be differing views. Indeed, that’s the point of gathering a diversity of perspectives. In theory, the government got what it asked for with expert reports that frequently point to uncomfortable questions. However, many disappear in the government summary that smooths over urgency and re-frames hard choices as balanced policy choices.
I was skeptical about the expert advisory group that Minister Solomon pulled together, fearing that it was missing too many voices and was given too little time to do its job. It would still have benefited from some additional perspectives, but the resulting reports suggest that the experts took their mandate seriously and provided candid, action-oriented advice on developing a national AI strategy. After digging into the what we heard report, I’m not convinced the government wants to hear it.
