MONTREAL — Innovation Minister François-Philippe Champagne knows he’s not the only policymaker around the world looking to create guardrails for all the new artificial intelligence systems popping up everywhere. “AI knows no boundaries,” he told The Logic, on the sidelines of All In, a globally-minded AI conference in Montreal. But “you need to start at home in order to influence.”

At the event Wednesday morning, Champagne unveiled a voluntary code of conduct for generative AI systems, committing participating companies to safety measures, testing and disclosures. Prominent Canadian scale-ups Cohere, Coveo and Ada have signed on, as have public firms OpenText and BlackBerry. 

Champagne is also pushing a new Artificial Intelligence and Data Act that would be mandatory for all firms, which Parliament is currently considering as part of Bill C-27. Meanwhile, Canada is working with the U.S. and EU on a separate, opt-in initiative designed as a stopgap ahead of binding laws. Both are also rolling out their own domestic measures. British Prime Minister Rishi Sunak is hosting an AI Safety Summit in November, which Champagne plans to attend. And governments are also looking to shape AI policy via the G7 and UN. 

That’s a lot of international initiatives. Champagne says they can all fit together, and that Ottawa can play a lead role. “A number of countries in the world [are] looking at what’s being done in Canada,” Champagne said, citing Ottawa’s work to establish the Global Partnership on Artificial Intelligence, a multinational body that gets policymakers, industry and academics talking together about the field. The new code is “a natural extension.”

Notably, Canada’s largest tech company hasn’t signed up. On Wednesday morning, Shopify CEO Tobi Lütke tweeted that he would not support the code. The firm has launched features based on generative systems. 

In an interview with The Logic, Champagne addressed Lütke’s criticism of his approach to AI regulation, laid out how he thinks the code fits into worldwide efforts to set rules for the technology, and explained why safeguards are a necessary condition for innovation in the field. 

This interview has been edited for length and clarity.

You have a code now. The White House has a code now. You’re working on a code with the EU and the U.S. There’s the G7 Hiroshima process. Are all of you balkanizing the regulation of AI?

No, I would think it’s all complementary. There’s no surprise about what we’re doing, because I [talk] to my G7 colleagues. A number of us believe that before we have full laws or regulation, we need a stopgap measure. The evolution of AI is going at lightning speed. 

I was urging my colleagues yesterday to pass [Bill C-27] as quickly as possible, because we want to lead nationally [and] internationally. Being the first country, I believe, in the world [with] a framework to regulate artificial intelligence is a big step forward. We need to rebuild trust in order to allow for responsible innovation, and we need to move from fear to opportunity. But before that, citizens need to know that the concerns that have been expressed by leading experts around the world have been heard, and that we’ve been acting. 

I think we are in a world where we need to have a number of frameworks in place. For me, this was the Canadian answer, to say: Before we have a law on the books, those who want to subscribe to a code on the principle of responsible AI, making sure it’s transparent and giving confidence to consumers, will have a competitive advantage [over] the others. That’s why it’s open. That’s why it’s voluntary. And that’s why I expect more companies to sign up.

How is the code with the U.S. and EU different from this one?

There will be discussion between us, the European Union, the United States. We wanted to have something concrete. There’s almost like a pre-ChatGPT world and post-ChatGPT world—the evolution and the speed of [AI] is requiring us to respond. We introduced the law to govern AI and data privacy in June of last year. Whilst we’re having discussions with our colleagues around the world, having a voluntary code of conduct is a way to say: Companies who sign up will use the principle of caution, to build trust and to allow responsible innovation. If we come up with something with our international colleagues, I think this will be complementary. We’ll continue to work with our G7 partners.

So it’s still an ‘if’ we come up with something with our international partners.

There’s discussions but we did not want to wait. We’ve been part of the process since day one. We need to do things in the meantime. But then there’ll be more to come.

You have a bunch of major Canadian AI companies signing today. The U.S. had Microsoft, Amazon, Open AI and other big U.S. companies—signing their equivalent code. Is every country going to get their domestic companies to sign on to their code? Or do you also want to see Microsoft and Amazon at the table?

It’s an open code. We started with the Canadian companies, the big ones like OpenText, Ada, Cohere, Coveo. There’s many others who want to sign up, [but needed] time to come to the table. Listen, if there are companies which are based outside of Canada but have operations in Canada that want to sign the code, that’s good.  

I want to ask you about a specific response this morning to the code. Tobi Lütke, the CEO of Shopify, tweeted about it. Have you seen the tweet? 

No. 

I’ll read it out to you. Yann LeCun, the AI scientist at Facebook who’s the third godfather of AI, used this phrase, ‘existential fatalistic risk from AI delusion disease.’ Tobi Lütke  called your code of conduct another case of ‘EFRAID’ and said he won’t support it. This is the CEO of the largest tech company in the country, which is adopting AI. What do you think of his position on this? 

I like Tobi. Let me start with that. I think he’s an amazing entrepreneur, an amazing leader in the tech sector in our country. But my job is to protect the public interest. 

This is voluntary. If he thinks that to promote his interests, he doesn’t need to sign the code, that’s a decision for him to take. I respect that. On the other hand, there’s a number of voices out there that are calling for framework to be able to operate. You mentioned what President Biden did in the United States, with some of the largest tech companies in the world. 

It is in Canada’s best interest, the best interest of companies, to be able to say that they will adhere to some basic principles on a voluntary basis that will allow for responsible innovation. I’m always happy to hear different voices. But it seems to me that there’s a large consensus in the public, if you look at the polls and what’s being done internationally.

I want Canada to lead. I don’t want to be at the end of the process. The voluntary code is a step forward for those who want to say to their customers and to the world that they’re going to abide by a framework. And others can decide what’s best for them. 

I’ve heard from AI entrepreneurs who say, ‘The government’s focusing a lot on risk, and they’re not focusing enough on promoting this technology and what it can do.’ Do you think the balance of your policy has shifted too much to the regulatory or risk side?

We need to build trust to allow for innovation. There’s a lot of concerns, and that’s why we addressed them with a new regulatory framework. But at the same time, I’ve always been there to say this has enormous opportunity for the economy. Think about productivity. Think about all the benefits for small- and medium-sized businesses. I’m the one advocating for that.