Canada can lead on AI by making trust a competitive advantage

Darryl White is the chief executive of BMO Financial Group.

Thirty-two years ago, one of my first investment banking assignments was to prepare a simple share-price comparison for a client.

I walked down the hall to our one central market-data terminal carrying a floppy disk, downloaded the information onto the disk, walked back to my workstation, uploaded the data, built a chart in Lotus 1-2-3 and proudly delivered it to my boss.

He studied it for three seconds and asked: “How would a different time period look instead?”

Back down the hall I went.

Canada joins OpenAI, Anthropic in AI standards pledge

Today, artificial intelligence completes that task – which took hours – in seconds.

But the lesson of the time stayed with me: The first answer is rarely the final one. AI can give us first answers with astonishing speed and confidence. Curiosity and human judgment determine if it’s the right one.

That lesson matters more than ever.

We already see why judgment matters in our daily lives. AI can help us plan, organize and work more productively. But we also need to be able to verify what it tells us, what information is not safe to share – such as banking passwords – and when a human decision must be just that, made by a human. AI advances are also driving legitimate questions around security, privacy, fairness and accountability.

“Responsible AI” doesn’t mean allowing prudence to become inertia. Instead, we should consider that responsibility promotes trust. Trust enables adoption. And adoption turns innovation into impact, which can help Canada escape the negative cycle of early invention and late commercialization.

The debate over AI safety shouldn’t be innovation versus regulation. We need both, in balance, with innovation enabled and accountability delivered.

The latest guidance from the Financial Stability Board, the international body that monitors the global financial system and promotes its stability, offers a useful model. Its new practices for financial institutions and AI emphasize organization-wide governance and managing risks throughout AI’s life cycle, encouraging boards and senior management to align AI adoption with business strategy and risk appetite.

That gives us a good place to start to address the implications of AI in the wider world. What are the risks? How are outcomes monitored? And who is accountable?

Answering those questions requires a principled response.

First, proportionate regulation. Not everything we use AI for carries the same risk. Leveraging a tool to summarize information should not require the same governance as one that could affect access to credit, health care systems or public safety. The greater the risk, the greater the requirements should be for testing and monitoring.

Social media has taught us the cost of waiting. These platforms have burrowed themselves into how billions of people communicate and consume information, and even into the foundations of our democracies. We permitted all this before we adequately confronted the consequences – including the threats posed to our most vulnerable population of all: Our children.

Only now are governments trying to establish guardrails, long after the technology has become nearly impossible to disentangle from daily life. We should not repeat that mistake with AI.

A second principle is international alignment.

Countries around the world have been here before, with successive races to reinforce global financial sector regulation co-ordination coming only after crises forced it.

AI presents the same challenge – and the chance for global action before a crisis, not after.

Canada’s recent international AI safety commitments are a promising step. Like financial sector regulation, effective governance of AI must ultimately work across institutions and borders. Financial institutions aren’t the ones developing the frontier models introducing the newest risks – those risks are managed today by research labs and software companies themselves.

Scientists push for creation of international body to regulate artificial intelligence

Today, the rules governing nuclear weapons aren’t local, nor are the regulations protecting the ozone layer, the laws of the sea and humanitarian protections. Technologies that carry global risks should have global rules.

Canada is already moving in this direction and can lead with a principles-based, risk-sensitive approach to AI governance that can adapt as the technology evolves. That builds on our tradition in financial services of establishing clear principles, strong accountability and balanced supervision, while giving institutions flexibility in how they manage risk and serve clients. Prime Minister Mark Carney’s call for a global technology stability board already embraces this approach, modelling it after hard-learned lessons from the financial sector.

Canada has real strengths. World-class AI research and talent, a growing technology ecosystem, abundant energy and the critical minerals that underpin the digital economy – after all, the digital cloud is made of metal. The opportunity before us is to combine these advantages with the innovation at hand, while ensuring that trust and accountability keep pace.

AI may increasingly give us the first answer, but human judgment, accountability and trust will determine the final one. If we can get the balance right, Canada won’t be a mere participant in the AI revolution, we will shape how the world governs it, with trust as our competitive advantage.