The new economics of AI

A decade ago, I was part of a deep discussion led by Ajay Agrawal and Tiff Macklem , on the potential of artificial intelligence. A few old-school executives left the meeting saying it was all overhyped. Macklem, who was dean of the 加拿大多伦多大学 - 罗特曼管理学院 , saw otherwise, as did Agrawal, his star professor and founder of the Creative Destruction Lab .

Last week, I was part of a similar discussion that Agrawal and Macklem helped lead, and there’s no questions AI — and Generative AI — is now a big thing. A really big thing.

It’s hard to find a sector that isn’t in the crosshairs of generative AI — the form that aims for algorithms to think and act like humans, and not just predict where the data’s going. But for all the technical progress — and stock market enthusiasm — there’s been less progress in the real economy, where the transformations will matter most.

Some of North America’s leading business economists came together in Toronto to figure out that gap, between theory and practice. Here’s what I took away from the Economics of Artificial Intelligence conference organized by the National Bureau of Economic Research :?

1. A new space race is advancing and will pressure governments to continue to think through priorities. Among the big choices coming at us: pressure for less regulation. Will precaution or innovation be the guiding principle?

2. We may need some space-race type agencies to lead us deeper into the Gen AI era. The U.S. already has some, just in less visible forms than NASA, and allies like Canada will need to catch up.

3. Governments need to adopt AI much more ambitiously. It can help with Ottawa’s current focus on public sector productivity, and be a market signal that Canada is in the market for services and solutions at scale.

4. We need to train and retrain lots more people for the Gen AI era, from software engineers to front-line managers and workers. The timing is ripe as Silicon Valley is in the midst of retrenchment, cutting hundreds of thousands of roles. Let’s not let that talent fade from view.?

5. We’ve entered a more shock-prone economy, with wars, pandemics and trade battles disrupting ?market efficiencies. That makes AI-driven economic modelling more difficult, and will make human intelligence — and wisdom — more valuable as more political, social and climate shocks come at us.

6. Citizenship — the values that bind us — will be more important than ever, as AI systems continue to Balkanize us based on our digital habits and preferences. Oddly, knowledge is fragmenting humans when it used to unite us, meaning that old algorithm called judgment will be needed more than ever.

7. Watch out for China. I’m of the view that that U.S. has a significant lead in advanced AI, largely because of its unique combination of venture capital, government research and academic excellence. But I was struck by the advances being seen in Chinese industry, where there’s a different advantage in the form of massive data sets and the cooperation (coercion) of consumers and industry.?

Macklem, who is now Bank of Canada governor, outlined how AI is already impacting financial markets, and could do so even more. It will not only change labour markets, but alter our ability to analyze and forecast those markets, as well as prices. AI should also enhance overall job creation, although we’re not yet sure how. One important fact: over the past decade, the number of digital jobs doubled while overall employment increased only 17%. Technology leads to opportunity, in other words.

But Macklem’s overarching caution is critical: “I’d be wary of anyone who says they know where AI can take us.”

Machines can predict a lot. So can humans. But even when we feel we can see the future, we’re still struggling to predict it.