How to make AI work for Canada 

There’s a paradox at the heart of Canada’s artificial intelligence (AI) story: Canada does a very good job at research and development of new technologies, but is poor at adopting them.  

 A panel of experts convened for a PPF member event – AI and Canada’s Productivity Puzzle: Cracking the code to unlock investment – heard that, despite Canada being home to some of the most respected generative AI research hubs in the world, adoption isn’t keeping pace. And, at a critical political and economic time, the fate of Canada’s true productivity potential hangs in the balance.  

 But how does Canada boost AI adoption? As it turns out, the barriers to AI adoption are more than structural. To harness AI’s true potential, Canada must overcome an inherent lack of trust, leadership and collaboration — that is to say, very human factors.  

Underpinning everything, however, is regulation. Canada’s ongoing lack of AI regulations perpetuates uncertainty and ambiguity and slows adoption over worries about compliance and best practices. Bill C-27, a broad AI bill that had been making its way through parliament is now dead thanks to parliament’s prorogation. While the federal government has outlined guidelines for generative AI use in the public sector and has said it supports a private industry-led voluntary code of conduct on responsible development and management of generative AI, there remains little clarity at a national level. 

In developing regulations it’s key that Canada does so in a way that spurs further AI adoption. Panelists suggested that, for one thing, Canada needs to find some way to standardize regulations across jurisdictions — an especially crucial issue for companies that operate globally to help simplify compliance. (The event was conducted under Chatham House rule, so attendee names are not being included here.) 

They would also need to be flexible. The panel heard that Canada should consider a tiered, risk-based approach, like that taken by the EU, which recognizes that not all AI systems carry the same risk.  

Canadian regulation ought to also establish language harmonization, laying out clear terminology that industry can understand, accept and use.  

Finally, Canada has to ensure that ethical design, fairness and inclusion are baked into systems from the start. As one panelist noted, AI requires large volumes of data, so compliance with Canada’s privacy laws and ethical frameworks is paramount to avoid bias and ensure fairness. Without these ethical standards in place for AI, organizations and individuals will be hesitant to adopt it. 

The panelists agreed that for Canada to fully leverage AI to fulfill its productivity promise, its culture needs to shift. Other investments will help, too. For example, Canada must put more into digital and physical infrastructure. The federal government announced last year that it would invest $2.4 billion in AI, the bulk of which was earmarked for computing capabilities and technological infrastructure, but the extent to which those funds will be put to use before a change in government leaves yet another gap in Canada’s AI puzzle.  

No matter the fate of that funding, governments must continue to encourage other pathways to adoption, including public and private partnerships, collaboration with local businesses, accelerators, research organizations and chip manufacturers. As one panelist noted, all sectors — like energy, health care or financial services — should be looking for partnership opportunities with the government to test new ground for AI. Unless Canada does this work itself, it may find its AI sovereignty slipping away.  

As for Canadian workers, including those in leadership positions, Canada must push for upskilling programs geared around AI. This cultural shift might be one of the more difficult for Canada to make. As one panelist noted, Canada is currently missing a digital mindset — meaning that innovators struggle from a lack of skilled resources to adopt new technologies. But workforce skills development programs will be crucial, as will understanding the nuances that come along with AI use in the workplace.  

For instance, as one panelist noted, AI has been found to boost skills for lower- or medium-skilled workers, but if it’s not implemented correctly, it can actually be a drag on highly-skilled workers (eg. a coder who no longer does their own work, but instead relies on prompts to an AI co-pilot). Businesses should be prepared for this uneven productivity impact and recognize that, even if AI can automate some portion of daily tasks, there are others that AI simply can’t replace.  

AI — and especially generative AI — is transformational technology. But how it transforms Canada is still very much in question. If the right pieces of the puzzle are arranged correctly, the potential for future productivity gains that AI helps to create is compelling. It will take a lot of work to get this right. As good as the technology is, it can’t create change on its own. That will be up to Canadians, overcoming barriers together.  

The Public Policy Forum thanks Deloitte’s Future of Canada Centre for supporting the PPF member event, AI and Canada’s Productivity Puzzle: Cracking the code to unlock investment. 

Join our network of policy leaders across Canada. Become a PPF member.