Evan Solomon Is the Opposite of an AI Doomer

It’s hard to know how to feel about AI when its capabilities advance seemingly by the hour. One day, it’s solving Canadians’ online-banking snafus and patiently robo-splaining the origins of our anxious attachment styles, and the next, we’re dating the chatbots themselves. AI monitors our real-time forest-fire emissions while its data centres furiously guzzle our water. It disrupts our jobs and, every once in a while, creates them: just ask Evan Solomon, Canada’s very first minister of AI and digital innovation.

A few months in, Solomon—a veteran journalist whose previous political involvement was hosting CBC’s Power and Politics—is still in the honeymoon phase of his portfolio. He’s trumpeting the economic upsides of AI to the country’s private sector and calling this our “Gutenberg moment,” a nod to the intellectual superbloom brought on by the printing press. Whether it turns out to be more of a Pandora moment for the rest of us is, at least partly, on him.

You were based in New York before you announced your MP campaign on LinkedIn a few months ago. During the race, you lived out of two suitcases and crashed on couches. Are you back-back in Toronto now?

I was working as a publisher at GZero Media when Justin Trudeau stepped down. Mr. Carney called and said, “We’re in a crisis. I’m going to run. I’d like you to run if you can.” My wife looked at me and said, “The kids are in university. We should do it.” While campaigning, I stayed at my brother’s house, then my buddy’s—now we’re looking for a place. My family’s here, so I know the city. But it’s like: welcome to the real estate market in Toronto! We’re renting and searching for a permanent home, like everybody else here.

That Carney call has strong superhero overtones. Captain Canada taps you on the shoulder and says, “We need you, Evan.”

Living in New York, you have a front row seat to what this crisis is about. There’s no default toward democracy; it requires a fight. Months ago, I was at my constituency office and met a volunteer—an Asian-Canadian man in his 60s—who was a pathologist at SickKids Hospital. It was a Friday at 10:30 p.m.—he must’ve been exhausted. But Canadians want to connect and defend this country. 

You and Mark Carney used to jog together when you lived in Ottawa. Any plans to resurrect that outdoor-run club?

I live in Toronto now, so…. We ran the London Marathon together, the Ottawa marathon—it was just a group of guys who love running.

AI has more or less become a race against Trump, who literally said he intends to “win the AI race.” How does technology factor into Canada’s big nation-building push?

This isn’t a race against one particular person. Two things are happening at the exact same time: a generational political realignment around the world and a technological disruption. I call AI our Gutenberg moment, similar to the way the printing press led to the Protestant Reformation.

How can AI secure our sovereignty, though? We’re heavily reliant on Americans for social media alone.

You need the basis, and Canada has a strong one. Two of the three godfathers of AI, Geoffrey Hinton and Yoshua Bengio, are here. We have Cohere, which develops large language models and AI products for businesses. (RBC uses them to analyze its stocks and generate reports, for example.) Of course, the U.S. has the big “hyperscaler” companies, but we’re helping ours commercialize. Canada’s also got the climate for data centres, because cooling is a huge issue. We just have to make sure they’re sovereign—not like, “Take our power, foreign company! And here are your microchips!”

When recently asked to namecheck a few Canadian firms putting AI to cool uses, you mentioned Vancouver’s GlüxKind. They’re developing a self-driving baby carriage. I have mixed feelings about that!

I met the GlüxKind guys in Paris at VivaTech, the largest startup and tech conference in Europe. Canada was country of the year, and about 170 of our companies were there. Emmanuel Macron comes in—it’s mayhem, right? Four days, 180,000 people. It felt like a rock concert, like when I was back at Shift magazine doing internet conferences 20 years ago. Every booth had another idea: AI that maps malls, robotics, crazy things. Among them was the baby carriage. It has sensors that read where traffic is if you don’t want to push it—the Europeans love it.

Sounds about right.

I know it’s like, “Do I need an AI-powered baby carriage?” But I remember when my dad was alive and automatic car starters came out. He was like, “Oh, you’re too lazy to turn a key?” In the winter? Who doesn’t want those now? With ATMs, Dad said, “Am I really going to put my money into the wall?” And then, “Am I really buying things by putting my credit card information on the internet?” Yes! AI is just a tool. If you’re a working carpenter and one day I said, “You can cut that with this electric saw much quicker,” and you said, “I prefer the handsaw!”… Okay!

A big priority of yours is getting the private sector on board with AI. This year, BenchSci, a biomedical tech company in Toronto, laid off almost a quarter of its workforce because it found AI efficiencies. One “problem” companies seem to be solving with AI is paying wages. How are you planning to protect the average Canadian worker amid all the excitement around private profit?

I’m glad you raised that, but we’ve got to be careful about not just fighting the last war. Of course you’re going to see moments where tech is used instead of workers, but mostly what you see is new opportunities. There are companies with valuations of over $1 billion in lofts in downtown Toronto—and you’ve never heard of them. It’s a real buccaneer’s moment, of people betting the farm. We can tell that story, as opposed to “something new’s here, and it’s a eulogy, not a celebration.”

I understand, but tech can generally be implemented faster than people can be upskilled. Nick Frosst, Cohere’s co-founder, worries AI could create more precarious jobs and worsen income inequality. This is the guy making this technology.

First of all, let’s be careful. There’s no consensus that AI is going to displace mass numbers of jobs.

But it will, right? No matter which research you look at, something like half of all jobs are now at least disruptable by AI. It will.

What I don’t want to do is play the game where you say, “Here’s a CEO that laid someone off,” and I say, “Here’s a company that has just hired 300 employees because of AI.” We’ll just go back and forth—new stuff, old stuff. Why do some societies thrive during moments of innovation and others don’t? How do we make sure ours does? We protect our education system, as well as citizens’ privacy and data, so people trust the tools. They have to see that they’re improving our system. Look at health care: Reveal Surgical, a Canadian company, developed an AI-powered detection system that can help surgeons identify cancerous tissue in real-time.

Your education involved plenty of religious studies. What does our borderline-religious obsession with AI say about us?

This is a continuation of the human curiosity journey. How do you communicate better? The Sumerians invented cuneiform. How about building monuments? Egyptians engineer things to move big rocks. What we’re doing isn’t new: building better tools to accomplish bigger things. But better tools often have unanticipated consequences. Sometimes the club you use to kill a bear might become the club you use to kill your neighbour.

Or the neighbour picks up the club and hits you.

Right. Or your neighbour goes and gets a crossbow.

My personal relationship with AI is basically ChatGPT on occasion, Gemini when my finger slips in Gmail. What’s your level of dependence in your free time?

I used AI today. I had to get a briefing on Bill C-27, a piece of legislation from a few years ago that has to do with privacy and data. I uploaded it to Google NotebookLM, asked the software to create a podcast and listened to it on the 15-minute drive to my constituency office.  Let me just see if it’s still on my phone. It’s amazing when you hear it. I’ll connect to a speaker.

…Is it working?

It takes a minute. I’ve sent that to tons of my staff. It’s pretty good, right?

It even injected a casual “let’s face it…” Like your chill, simulated best friend.

Oh, yeah. They completely modelled it on every PBS or NPR show. No hallucinations, either.

A Harvard Business Review study recently showed that the No. 1 use for ChatGPT is therapy and companionship. Have you ever used it for that—or know anyone who has?

I just spoke to a doctor about this. If you ask a therapist if they’ve used a doll or a hand puppet, they’ll go, “Yeah. It’s a useful therapeutic tool.”

Anything’s better than a puppet, in my opinion.

Everyone’s asking: how do we combat the loneliness epidemic? Social circles matter, and you’re lucky if you have a partner and kids and friends. A lot of people don’t. I had elderly family members whose only companion was a TV. I’ve met constituents whose lights were all off to save electricity but who left the TV on for the same reason. If you had something to carry on a conversation with—and not just watch—that might be better. But there was also that Joaquin Phoenix movie, Her, where he was in love with an AI. Of course we have to be vigilant.

With two university-aged kids, you’re clearly past the “should we just give them the iPad” issue. How do you talk to them about balancing the AI world with the real one?

Luckily, my kids haven’t had a massive screen-to-nature imbalance. My son was a guide at canoe camp, and my daughter was a tree-planter. (So was I, actually.) I’ve told them to try to get comfortable with AI, so they’re not swayed by three articles. Think about the billions of dollars and hours teachers have spent trying to make sure every kid can read and use a pen to scribble the alphabet. Spelling may be less important now because of spellcheck, but there are still going to be new things to learn. 

RIP, essays.

If we’re typing less and dictating to speech-to-text tools more, the ability to speak clearly becomes more important. You don’t want to have to go through a transcript of me being like, “Oh, like, hey Katie, like, this is a great interview, like—oh no. Go back. Oh god, like…” You want: “Thanks so much for this interview, Katie. If you have any more questions, let me know.”

This interview has been edited for length and clarity.