The Decibel - Carney’s plan for AI in Canada, explained

Episode Date: June 8, 2026

The federal government released its plan for AI, and it includes a lot of new money ($2.3 billion, to be precise) towards training, adoption and supporting Canadian businesses and tech companies. There was a lot of anticipation leading up to this strategy’s release, both from those excited about how Canada plans to build and harness the technology, but also from those who are concerned about the potential ramifications and safety of AI. Joe Castaldo is a staff reporter writing about AI for The Globe. He joins us to walk us through what’s in this strategy, and what’s not. Questions? Comments? Ideas? Email us at [e-mail] Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

After months of anticipation, the federal government's plan for AI is here. The question isn't whether AI will transform our lives. It will. AI is already changing how we work, how we learn, and how we connect. The question, the question is, will it improve the lives of all Canadians, or benefit only a few? Prime Minister Mark Carney announced the AI strategy on Thursday with AI Minister Evan Solomon. It includes a few big things we're going to talk about, like how the government plans to get companies to adopt AI, how they plan to help Canadians better use and understand it, and the money they're putting towards keeping tech companies in Canada. We'll stress test some of the details, like whether AI adoption will produce 250,000 jobs in Canada,

and what the government is planning to do to ensure our safety around the technology. Joe Castaldo is a staff reporter for the globe who writes, about artificial intelligence. Today, he walks us through this strategy, what's in it and what's not. I'm Cheryl Sutherland, and this is the decibel from the Globe and Mail. Hi, Joe, nice to see you. Nice to see you. Okay, so we finally have this AI strategy from the federal government. What is the government doing here? Like, what is an AI strategy? Oh, boy. So at the highest, highest level, it's like a recognition by government that, AI is a powerful technology that has a potential to change society, change the economy for better

or worse, probably a bit of both. And it's their attempt to get their arms around it and figure out how can we use AI to better the lives of Canadians. And I think the setting for the announcement actually is kind of notable here. So Mark Carney announced it at Toronto General Hospital. with a backdrop of like healthcare workers and some medical equipment. And I think that's to try and reinforce the point that they want to make, that AI can be a force for good, a lot of health care applications. And that's kind of a message throughout this strategy.

Okay. And I think something else the government is trying to do here is about trust, right? Like they're trying to get us to trust this technology? Yeah. So the government talks about six pillars of the strategy. but also three like principles, so it all gets mixed up. But like to summarize, trust is a big one. Canadians have low trust in AI.

The government needs Canadians to trust AI so that they will adopt AI. And then if they adopt AI, that will help the country, help the economy. So and all of that leads to Canada having a role in shaping the development of AI. So all of that ties together in some kind of way. in this strategy. And so where does Canada stand globally in creating an AI strategy? Like have other countries done this as well? Yes, yeah.

The UK has. There's a European Union strategy. I think Japan and Singapore. Canada prides itself on being early to AI strategies. Like it really goes back to 2017. There was a pan-Canadian AI strategy that was really research-focused. So it helped to create, you know, MILA, in Quebec, the Vector Institute, Toronto, Amy, and Edmonton, just these research ecosystems.

There was another strategy of sorts in 2024 that had money for, like, towards data centers and compute infrastructure. So this is the latest evolution of that. And I think arguably trying to be the most comprehensive. I know I've asked you all these questions about what the strategy is and what they're trying to do. But I do want to ask you one more before we get into the details about the balance here. What is the government trying to balance with this strategy? So it's interesting. There's always some cognitive dissonance when you talk about AI.

And I think that comes through in the strategy as well. So like the first pillar is about protecting Canadians from the harms of AI. And yet, you know, the bulk of it is about adoption and how we need to use AI. So they recognize that Canadians have a lot of concerns about AI. I think opinion polls show Canadians are fairly mixed. Like, you know, some think it'll be good for society, some think it'll be bad, and then there's another portion that are uncertain.

And, you know, the concerns run the gamut from, you know, job loss to misinformation, to students aren't going to be able to learn and think for themselves. Environmental concerns around data centers is a whole long list of things that Canadians are worried about. So they have to try to address those concerns. while convincing Canadians to use AI and Canadian businesses to use AI because, you know, it could theoretically help the economy. So, I mean, the strategy even notes that half of Canadians think AI is a threat to humanity. Yeah. Yeah.

Yeah. It's a pretty extreme statement in a lot of ways. So it's a tricky balance here in this strategy. And leading up to it, you know, Minister Evan Solomon has been saying, you know, there's AI cheerleaders and there's people with pitchfork. who want to stop development, you know, we're going to be very pragmatic. They're trying to position it as like straight down the middle. I think the framing that he's using is a bit disingenuous. I mean, people with pitchfords, let's say, they have legitimate concerns about AI that need to be addressed. So they're trying to thread a needle.

Yeah. Let's start with the adoption side of what the government is doing then, and we'll come to the concerns. What is the government doing to help AI companies in Canada scale up? So this is actually where a lot of the spending is. There's two big elements to it. There's $700 million for an existing program that basically helps to cover the cost of building, running, AI models for Canadian companies. That's important to developing AI. It can be expensive to pay all these computer processing costs.

So it's, you know, a subsidy effectively. I don't think companies are going to complain about that. it's money that they can access. I think what's interesting about it, though, this program was first announced two years ago in 2024 with $300 million. This is the AI compute access fund, right? Yes, that's right. It took a year or so to open up to applications.

And then the government only made the first disbursements in May. So it's like two years for money to flow. And so it did get a lot of criticism that this is moving way too slowly, especially given how fast AI is moving. And Minister Solomon has promised, you know, we're going to move faster this time around. So, you know, we'll wait to see about that. The other component is a new fund called the Canadian Tech Growth Fund. It's $500 million to invest in what the government says will be the most promising AI companies. What is a promising AI company?

Do we know? We don't know. We don't know how they're going to assess that. But it's designed to address a funding gap where if companies grow to a certain size, they have difficulty raising funds in Canada and maybe they have to turn to foreign investors. So it's an attempt to address that and ensure that companies can remain in Canada. Again, we'll see how this plays out. There's a lot of factors involved in sort of why an entrepreneur might leave for the U.S., which is happening quite a lot. lot, which we've reported on. Again, you know, how quickly can this get up and running? Like, it's, if you talk to entrepreneurs, it's just so easy to raise money in the U.S. It happens so fast. That's why they go there. There's more customers in the U.S. for AI products. You know, this doesn't really address that. Remains to be seen if it will achieve that goal. What about regular Canadian companies? Like, how is the government approaching AI adoption across the business sector? So this is kind of another big theme. or goal in the strategy is AI adoption in the business sector. It's low, generally. The statistic that gets quoted a lot is 12% of Canadian companies using AI essentially. It's even lower with small and medium-sized businesses. Is that low compared to other countries? Yeah, generally. Yeah. And the goal is to get business adoption up to 60% by 2034. So it's a goal. And

To do that, they have some money behind it. So another $500 million for regional development agencies to help companies with AI adoption and like commercializing AI. They reference a BDC program that already exists that does something similar. So, you know, what these programs are is if you're a company and you are solving a specific problem with AI, you can get, you know, a loan for it to help develop this. that's kind of one prong they're using to boost adoption. They've also identified priority sectors in Canada, where they think, like, Canada can really win in, like, AI for manufacturing or AI for agriculture. And, you know, we're going to bring all of these parties together, companies, government, researchers,

and we're going to attack the hardest problems in these sectors. And that will be, they call it the engine of Canada's AI adoption. which is so vague. Yeah, I was going to say it sounds interesting, but not really much detail there. No, and that's, you know, I think one of the takeaways from the strategy is there's not a lot of detail around a lot of this. So, you know, that part of it will have to see. It's ambitious. I don't know the mechanisms to really drive adoption there. And like adoption is just kind of a headline number.

Like if a company gets a Microsoft co-pilot subscription, is that AI adoption? And that doesn't tell you, like, are they getting any value of it? Is it out of it? Is it really moving the needle in any substantial way? So we'll see how it shakes out. How about on an individual level? Like, what does a strategy say about helping Canadians get up to speed on how to use AI? They say they will be launching essentially like a national AI literacy campaign to provide free AI education to every,

every Canadian. And they frame literacy as, you know, understanding AI, how it works, how to assess, you know, chatbot responses, what are the limitations, and then, you know, applying it in your life or at your business, so you're getting something out of it. What this course, this training material looks like, we don't know. They talk about, you know, using public libraries and community organizations, you know, to help get this out on that. There are a lot of free resources out there already about how to use AI. They're just not government-sanctioned, I suppose. The goal is to reach everyone with this.

So we're all well-versed on AI and can start using it. I understand that a lot of these literacy efforts are focused on young people, though, right? Can you tell me about that? Yeah, there's a focus on college and university students as well. And one of my favorite lines and the strategy is how Canada will provide every post-secondary education student with a trusted AI agent. Which is what? We do not know. Okay.

We don't know what this is. What will it do? How is it trusted? Why is it trusted? I think it's a little detail, but it really captures the vagueness and the lack of detail on a vote that kind of runs through the strategy. There's a technical briefing for journalists before they announced the strategy where you can talk to basically the bureaucrats behind the strategy and ask them questions. And this question was asked. It was the first question asked, as I recall. And specifics were not forthcoming.

We'll be right back. So in the strategy, the government says it'll create 250,000 jobs through AI adoption by 2031. That sounds like a lot. But can you explain how they arrived at that number? Well, to put it in context, 250,000 jobs. So job numbers were released on the day we're talking. Yeah. In May.

We created 80,000-something-something jobs? Yeah, 87,000. There you go. More than 87,000 jobs. So, you know, we're talking 250,000 over the next few years. So that's the context we're dealing with. But also, the strategy says it will be created through AI adoption and not much more than that. And so I did ask Evan Solomon, how did you arrive at the number? And he said they used

basically some modeling from the OECD that looks at job creation based on different adoption rates. And they use like the most aggressive adoption rate. And then this is how many jobs will be spit out. So I don't know how seriously to take that number because nobody knows how AI is going to impact the labor market. You know, what new jobs will be created? What jobs, you know, will not be as necessary or will be replaced. This is what economists, like, nobody knows. It's highly debatable. I also ask, like, did you do any modeling about job replacement or job loss? And he said, no. And he said, why? And he said, because it's too uncertain. I would argue, you know, job creation is just as uncertain. But, you know, the fact that they're leading with that

sort of tells you how they're thinking about this. I think that one thing that people are very concerned about when it comes to AI is, of course, job. loss, right? What do we know about potential job losses through AI integration? Is there anything in the document about protecting jobs at all? Well, I mean, it acknowledges certainly disruption to jobs. And that's if you, you know, ask the government about that, that's where they will bring up, you know, literacy and training and job opportunities for young people, right? So they have AI skills to exist and make a living in our AI-enabled future. That's their response to it. Certainly unions were looking for stronger protections, things like mandating companies provide

retraining to employees who are impacted by AI measures to prevent companies that receive public funding to replace jobs with AI, things like that. None of that is in the strategy. Yeah, because the fear there was that these companies could get the funding, but there's nothing in the, in this document saying anything about protecting jobs, right? More or less. Yeah. Yeah. It's more about retraining.

Skills and retraining is the way that the government plans to deal with labor disruption in this strategy. Like there's no more support for, you know, employment insurance and things like that or social safety net programs that that might help workers who are displaced. It's skills training. And even then, it's fairly. vague about how that will play out. One of the elements that was missing has to do with intellectual property protection. In the task force, there were some proposals around that.

So, you know, IP that is publicly funded through universities, for example, there should be some mechanisms to ensure it stays in Canada as opposed to getting acquired by a foreign company. That's been a pattern in Canada. I think, you know, most notably, Jeffrey Hinton, who's, you know, one of the pioneers in AI, you know, long-time University of Toronto professor, you know, when he hatched his startup out of the university, it was acquired by Google. And that has led to untold riches for Google. So there was little to no talk about IP reform. And that's important. A lot of people argue in the AI era. What does the government say about how it plans to handle the issue of protecting Canadians from harms related to AI? Yeah, I think that is important.

And it's important to the government, too, if you read the strategy, again, like trust is a big theme. Protecting Canadians is the very first pillar. What's ironic is there's not much new on that front. That's surprising, no? I feel like this is something that's probably on a lot of people's minds when it comes to AI. Surprising in a number of ways, too, because there was a public consultation leading up to this strategy. You know, they got 11,000 submissions or something like that. And I think regulation and oversight and guardrails and penalties for AI companies, all of that was a big theme running through the responses, which the government itself acknowledged in its summary.

But there's not much there. I mean, I wouldn't expect the government to introduce legislation as part of a strategy. But what they talk about is stuff that they've already talked about. Like they plan to introduce new privacy legislation and online harms. legislation. They've been talking about that for a while. We haven't seen it yet. Don't know when it's coming. And so it's just the language that we've heard before. And I think people were looking for just some more clarity around that. You know, what do they hope to achieve with this legislation? Just more of their thinking around that, at least. And there just, there wasn't anything new there.

I think, you know, when you talk to the government about this, they'll point out some other things like, oh, we have $50 million going to the Canadian AI Safety Institute, which is fine. I mean, that's academic research about AI models that takes a long time to translate into something tangible. But yeah, I think people were disappointed on that front. And I don't, and because of that, you know, it's unclear to me how this strategy achieves the government's goal of trust, right, if there aren't sort of stronger safeguards in place. Do we have any more information? I think you did say that we don't know when this is coming out,

but what did the AI minister say about when this legislation around online harms will come out? He'll only say it's forthcoming, which they've been saying for a while. Maybe just step back for one second. It's kind of interesting how thoughts on regulation have shifted. Like in 2023, when AI was really exploding, regulation was a big priority in Canada and elsewhere. And then it kind of fell by the wayside. Really, like last year in particular, after the election in the U.S. of Donald Trump, you know, Vice President J.D. Vance had this speech at a big AI conference in France where, you know, he basically said it's going to be pedal to the metal with AI development. And a lot of countries like took their cues from that. I think to some extent Canada included. You know, Evan Solomon came in saying he's not going to introduce AI specific legislation, just privacy. and, you know, data protection.

And then I think, like, the winds are shifting again more recently. In Canada, you know, Tumblr Ridge happened. So that's kicked off, you know, a lot more conversation about what do we do about chatbots and youth. Can you explain a little bit about just to remind people what happened there? So the shooter in Tumblr Ridge had discussed scenarios involving gun violence with chat GPT. In the months beforehand, these conversations were flagged. but OpenAI did not report this to law enforcement. And, you know, the interaction between AI companies and law enforcement had not been a big topic of discussion before then, right?

But it's just kicked off all these conversations again. So that's one element. And there's also all the other examples of, you know, we're hearing about kids that are taking their own lives, right, who have had conversations with chatbots. And there's concerns there as well. Yes. Yeah. And, you know, they're pursuing recourse through, you know, the courts. You know, there's conversations about, you know, banning youth access to chatbots.

We don't know which way the government's going to go. We don't even know if chatbots are going to be a component of the online harms bill, which, you know, had been focused more on social media, the discussion around it at least. So there's still a lot of questions. And the other component, I think, that's changed things is security. You know, there's these powerful AI models that can, you know, find software vulnerabilities. And I think that's scared a lot of policymakers into, you know, rethinking like, okay, how do we get ahead of this?

That's a myth those you're talking about. Yes, exactly. So the winds are shifting. I don't see it reflected in the strategy. Conservative MP, Melissa Lansman, spoke after the announcement and said this plan was missing a lot of details. I think today's announcement was a lot of fanfare, short on details and a lot of hollow words from a podium.

I think Canadians are expecting real answers on safety, on security, on, on privacy and on the future of AI in this country. Others have said the same thing. We've kind of talked about the vagueness of this plan. Is this satisfying to the people who have been waiting for the strategy? I think there are elements of it different constituencies will like. I mean, there's more money for the research ecosystem here. So researchers are going to like that. I think business groups in general like the goal of AI adoption.

I think some of them like the fact that there will be a new source of funding. But I think overall, it's sort of a typical liberal government kind of budget document in that it tries to do a lot of things. So a little bit here, a little bit there, and spread money around evenly, more or less, but doesn't take any like big bets or sort of reflect any radical new change in thinking in terms of how we've been doing things until now. Joe, thank you so much for coming back on the show. Really appreciate it. Thanks again. That was Joe Costaldo, who reports on artificial intelligence for the globe. That's it for today.

I'm Cheryl Sutherland. Ali Graham mixed this episode. Our associate producer and intern is Cynthia Jimenez. Our producers are Madeline White, Rachel Levy McLaughlin and Mihal Stein. Our editor is David Crosby. Adrian Chung is our senior producer, and Angela Pichenza is our executive editor. Thanks so much for listening.