There appears to be no respite for Canada’s economy with surging oil prices, supply chain bottlenecks and labor shortages all testing its resilience. Add to this Russia’s invasion of Ukraine and it’s clear the immediate future will be daunting.
Canada cannot fall back on high productivity to weather this storm. Where we were once roughly on par with the U.S., productivity growth averaged only 0.7 per cent per year in the 2000s, less than half its 1990s rate and almost only one-quarter of America’s 2.7 per cent annual growth rate in the same period. This is a long-standing challenge, which successive Liberal and Conservative governments have attempted (and failed) to tackle.
Though there are multiple structural factors at work, the Bank of Canada, the OECD and others blame Canada’s poor performance at least in part on a poor record of technology adoption. If Canada is to rebuild a more resilient, prosperous and competitive economy in the aftermath of the pandemic, it requires a more positive story about – and massive incentivization of – technological adoption at home.
Canada has great potential to harness the economic gains from the technology most likely to drive growth in the 21st century: artificial intelligence (AI). Like the steam engines that powered the industrial revolution in the 19th century, and electricity in the 20th century, AI is set to add exponentially to global GDP, with estimates as high as $15.7 trillion by 2030.
Canada is well-placed to capitalize on its early investments and advantages in AI research and talent to implement a second phase — one that focuses on encouraging wider adoption across the economy.
Canada has a head start in AI research and innovation
AI is here to stay, and governments must find a way to harness its potential to maximize positive gain. Controlling for certain specific risks is necessary and will help generate trust in the technology. But we should not conflate the risks posed by a minority of AI applications with the technology at large. Regulation can be done in a deliberate and targeted way to address high risk or sensitive applications of AI, without constraining the use of the technology in general.
Policymakers and politicians should expand their horizons when thinking and talking about AI to include its potential for innovation, wealth creation and employment augmentation.
The federal government does not appear to have an innovation policy strategy, let alone an AI innovation strategy. Instead, there’s a hodge-podge of different policy measures.
Monique Leroux, chair of the federal government’s Industry Strategy Council, told the Globe and Mail earlier this year: “We need to have a more robust, long-term vision about what we want to achieve …. You need to have a strategy.”
The latest federal budget brought a slew of new innovation-related investments. But critics have voiced concerns that there appears to be no strategy to ensure new programs will work together with existing ones.
Sadly, measures to date have been as ad hoc in AI policy as they’ve been in innovation policy writ large. Despite massive investments in fundamental research, there is no apparent policy roadmap across the technology readiness levels (TRL) scale and into large-scale adoption. This is a recurring problem of fundamental research successes.
The 2021 federal budget announced $443.8 million over 10 years to “renew the pan-Canadian AI strategy” – much of which will go to fundamental research support. But the 2022 budget mentioned a relatively small investment to maintain the AI supercluster.
Stay in the know with veteran reporter Kathryn May. Sign up for routine and out-of-the-ordinary news about the public service with The Functionary, our new newsletter.
Overall, this gives little confidence that AI adoption is of central concern, despite its standing as our technological path to productivity growth, and despite the fact that Canadian businesses are behind when it comes to AI adoption.
The next phase for Canadian AI
Canada should produce an ambitious AI industrial policy strategy, to boost productivity and create high-paying jobs.
Much of the untapped potential of AI lies not only in the development and scaling of AI startups but also in its adoption beyond the tech sector. Like the steam engine, electricity, semiconductors and the internet, AI is a general-purpose technology. It has potential across the entire economy. From financial services and agriculture to manufacturing and shipping, AI can be deployed at scale to lower costs and improve productivity.
We should take inspiration from the arrival of electricity in the late nineteenth century. Electricity transformed the technological and geopolitical landscape. It also fueled the American ability to leapfrog other nations’ industrial capacity.
As Stanford’s Jeffrey Ding argues, U.S. leadership depended less on invention. Instead, its capacity to adopt and scale applications of electricity across industries made the difference, from incandescent lighting to electric trams. The lesson is that the key institutional competencies are those that facilitate the diffusion of general-purpose technologies.
The next phase in Canada’s AI strategy should therefore spur the wider adoption of AI-enabled processes across the whole economy. This means not only in supply chains, and not only using the supercluster policy format. It will also require bold thinking and, first and foremost, a plan.
First, businesses need to be sold on the uses and benefits of AI, not just focused on the risks associated with a limited number of applications. Whatever regulatory approach is taken should be laser-focused on these risks, and provide certainty and assurance to businesses looking to invest in and use AI. Politicians should be reassuring on this front and should highlight the positive economic impacts of AI adoption whenever possible and appropriate.
Second, the federal Innovation Science and Economic Development (ISED) Department should complete a landscape analysis of policies that either directly or indirectly support AI adoption. This builds on the work done for the innovation and skills plan, and the commercialization working group report of the AI advisory council – now two-and-a-half years old.
Third, the Canadian government should “show rather than tell,” and aim to be the world’s leader in government adoption of AI. This could lead to smarter policymaking, reimagined service delivery and more efficient operations.
Fourth, where it is possible and appropriate to add AI adoption as a priority to an existing policy, the government should do exactly that. This could include dedicated AI funding for the Innovative Solutions Canada program. It could also mean the Canada digital adoption program or the recently announced Canada Growth Fund have specific programs for AI adoption.
These are necessarily brief and high-level recommendations, but in a sense that’s the point. There is no indication that the federal government has gone any deeper on AI adoption, nor have we seen true buy-in and support for AI adoption from responsible ministers.