The decision to create a Minister for Artificial Intelligence sends an unmistakable signal that the Carney government recognizes the need to prioritize AI as a core part of its economic strategy. My Globe and Mail op-ed notes that while few doubt the importance of AI, what the federal government should do about it is far less certain. The Trudeau government emphasized both government handouts and regulation, with billions in AI spending promises on the one hand and ill-considered legislation that was out of step with global trends on the other. The result was a mish-mash of incoherent policies that left the AI sector confused, civil society frustrated and Canada at risk of being left behind.
Elevating AI to a full ministerial position suggests Prime Minister Mark Carney wants to fix the status quo, but in some ways the new office looks like an impossible job dressed up in ambition. Evan Solomon, the minister, steps into a role full of symbolism but operationally murky. Mr. Solomon may well find that cutting more cheques or introducing regulations won’t solve the issue.
Under Mr. Solomon, will the government continue to view its AI role primarily through the prism of funding programs? There are certainly plenty of companies that will urge it to dole out hundreds of millions in support of computing infrastructure or business tech adoption. Yet recent failures involving “investing” billions in electric vehicle manufacturing or supporting e-commerce implementation highlight the risk of betting big on government spending as the solution.
Indeed, AI investment worldwide runs into the hundreds of billions of dollars, and the Canadian government contribution is never going to be more than a rounding error in total AI spending. That doesn’t mean there isn’t a role for government to help catalyze growth in the domestic AI sector, but if funding announcements become the hallmark of Mr. Solomon’s tenure, even ChatGPT will tell you the initiative is a failure.
Mr. Solomon also faces a quandary in determining which regulatory approach to pursue. The Trudeau government used the European Union’s AI Act as its north star, expediting proposed legislation that largely mirrored its approach. Yet the European hopes for a “Brussels effect,” in which countries worldwide adopt its policies, is rapidly turning into an anti-Brussels effect. The rush to regulate during the early stages of AI companies and technologies has begun to look ill-considered and damaging to those countries that still harbour hopes of AI economic leadership.
The EU AI Act’s biggest challenge is that the U.S. under President Donald Trump offers a far different vision, one in which economic growth and innovation are prioritized. In practical terms, this means restrictive regulations give way to more flexible guidelines. The U.S. approach will leave many privacy and AI experts concerned about the ineffectiveness of light-touch regulation, but it is likely to appeal to AI companies, investors and consumers more focused on access to AI services than the potential for privacy, security and equality risks.
This leaves Canada in a tough position. Follow the European model and run the risk of losing companies and capital to the U.S. Adopt the U.S. approach and face the prospect of real-world AI harms without an effective regulatory response. Ultimately, Mr. Solomon won’t be able to sit on the fence – Canada is going to have to pick a side.
Once the legislative approach is selected, one more question must be addressed: Who is really going to lead on AI policy? Mr. Solomon has the title, but AI responsibility has to date been spread across multiple departments. The Industry portfolio, now led by Mélanie Joly, has been the home of AI policy, as it introduced the combined privacy and AI bill that failed to pass through the House of Commons. Steven Guilbeault’s Canadian Heritage department plays an important role in copyright policy, which is among the most contentious AI issues, given the implications of scraping internet data for training models. Add in Public Safety, Justice and the Treasury Board and there are too many cooks in Canada’s AI policy kitchen.
Precedent suggests AI policy will gradually become a dreaded issue of “shared responsibility.” While there is nothing theoretically wrong with bringing multiple departments to the table to hash things out, the reality is that it often leads to long delays as bureaucrats work to find a policy consensus, which itself often seems to prioritize turf protection over effective policy. Mr. Solomon has the chance to forge ahead with a new approach, but only if he can make his mark by first asking the right questions.
The post Solomon’s Choice: Charting the Future of AI Policy in Canada appeared first on Michael Geist.