A general-purpose technology is one that is so revolutionary that it has impacts across everything.

It is suggested that AI is at the top end of that.

Today, countries are facing a numerous number of challenges. Where AI has a real opportunity is to tackle some of these challenges head on.

What it’s enabling is for smarter digital infrastructure to be developed at the very core. So, much like the Industrial Revolution came up with a new array of infrastructure developments, this is the rise of a very different smart digital infrastructure.

We still don’t know definitively of how big a deal AI is on things like economic productivity.

Productivity is one of these economist obsessions that no one else really understands

or even really cares about. But in the end, it is really important

because it measures how productive is the labour, and it matters, ultimately, in

terms of your economic wealth and growth.

That’s why this paper looks at both of those. Productivity as a core issue — how’s AI going to play into it? It’s going to matter big time.

The issue is that a lot of countries have certain capabilities and certain gaps in their existing infrastructure, which they need to address. So, in this paper, what we’ve done is introduce, essentially, a practical framework to understand how countries can translate

AI investments into possible productivity gains. And we’re calling these our core transmission conduits.

What’s important to note here is why we’re using this word, “conduit.” The idea is that it captures the idea of something that enables the flow from one point to another. And, in this case, the critical question is, how does AI actually lead to higher productivity at the national level?

And the answer is that it’s not automatic.

There’s no straight line from AI innovation to economic growth. Instead, productivity gains

will flow through these conduits, which are essentially mechanisms

that carry AI’s potential into real-world results.

These conduits were chosen based on how actionable they are, how measurable they are,

and the point that they’re all interdependent.

The first one is the technological capability. It’s the backbone, right, it covers the country’s technical readiness, essentially, to develop and deploy AI at scale.

The second one is [the] application and markets conduit. Here we’re really trying to understand and unpack, how AI is actually going to be used in the real economy. So, is there a demand side pull? Will companies adopt it? Will there be investments coming in?

And the last one, which really brings all of this together. What are the rules? What are the incentives for making all of this work? And that’s the policy and regulation conduit. And as we prepare our populations and our countries at scale for developing national AI strategies, this is where this conduit will play an important role in understanding what’s required.

Scenarios are very helpful when you’re dealing with a high area of uncertainty.

And if you think of it as a square, we looked at four scenarios and think of each one of those scenarios in a corner of that square. And the reality is playing out in the middle

as they’re kind of jostling between each other and becoming more plausible.

There’s no one answer to it, because it’s the future.

Number one is flat AI, which is stagnation over at least the medium term of where AI is going.

Number two, US-led AI, which is about friendshoring, bringing in allies together and

doing a lot of things closely together.

Third is multipolar AI, which is the idea of a thousand flowers blooming essentially in all main economies and AI really taking off.

The fourth was artificial general intelligence, which is an absolute revolution

in terms of the technology and the impacts that would be felt across

productivity, growth and labour.

They’re plausible outcomes or futures but they’re not probable necessarily, and you’re not assigning probabilities to them. You want the decision maker to look at all of these and then determine which ones they think are the most relevant, or even ones that they want to see and help drive towards those scenarios.

We’ve seen that different countries are at a different level in their AI productivity gains skill.

What’s clear from our paper is that no single country will succeed in isolation.

AI-driven productivity is not just about innovation and competition, it’s also about coordination. So, the comparative advantages or disadvantages a country has — they will have to complement and work with others to bridge those gaps.

And that’s where the real challenge for the global community lies, is can we create a governance architecture that moves away from just sheer competition to coordination?

As a global think tank, you can generate useful ideas in the spaces where there are a lot of things changing very quickly, where there aren’t established rules or governance around it.

CIGI has ideas on these areas where new technologies are converging with governance challenges. And then, in addition, we have a power to convene a global discussion on this, right, and even though it’s a G7 conversation, we engaged with global thinkers.

The challenges of AI at scale will be for all of humanity to face.

I don’t think any country in isolation will be able to tackle it on its own.