As businesses push ahead with ways to develop artificial intelligence, governments are looking at ways to regulate its impact. Kevin Hebner, Global Investment Strategist at TD Epoch, speaks with Kim Parlee about the state of regulation and the implications for the companies at the forefront of AI innovation.
Print Transcript
[AUDIO LOGO]
As businesses push ahead with ways to develop artificial intelligence, governments are looking at ways to regulate its impact. The EU is amongst the first, recently approving a major set of regulations. But not everyone agrees this could be the best approach. Joining us now is Kevin Hebner, Global Investment Strategist at TD Epoch, to tell us what he sees when he looks at the industry. Kevin, it's always a pleasure to have you here. Can we just start off with, how effective do you think some of the proposed legislation coming out of the EU is going to be in regulating AI?
I don't think it's going to be very effective. There are some good aspects of the regulation they have. For example, in terms of AI-generated content, it's going to have to tell you that this is from AI. It's not something that a person created. Or if you're talking to a chat bot, it's going to have to tell you you're talking to a chat bot, not a real person. So those seem like good things. But overall, they're trying to regulate this very broad technology.
It'd be like regulating electricity, or the computer, or internet rather than regulating specific uses of it. And if you try to address every particular potential harm from something as broad as AI, I think the legislation is being so broad, covering so many activities, that it's going to be ineffective. Moreover, AI is developing so rapidly, we don't really know where it's going to be 2, years, 5 years, or 10 years from now. And I think that's a big issue. And every time you have regulation that's premature, overly broad, it ends up doing a lot more harm than good.
We can talk-- there's definitely some different tacks, I would say. We know in the States, they're going a little more laissez-faire so far, I will say, in terms of what's coming out. We'll get to that in a second. But if we take a look at that came out of some-- from an EU regulator, they talked about how the world should maybe look to the EU for a basic model of how to balance freedom and safety, not to focus on just the technology, but on risks likely to accompany its various uses.
Yes.
What do you hear when you hear that?
Yeah, so that is the perspective of the key architect of the EU legislation. So very much focused on risk, potential harms, not paying nearly enough attention to the benefits, the innovation, that comes with AI. And I think this is a big worry. When you look at the EU, they're regulating themselves to last place. They have very few tech companies. There's lots of very good AI people from Europe. But most of them now work in the United States. A couple smaller companies in Europe, but relatively little.
And I think given that these big technology innovations, they're why we're rich, they're why we have jobs, why we're relatively well paid, I think it's a big concern. And in particular, it's going to harm relatively younger people in the EU. And they're not going to have the opportunities that people in Canada, the United States, and other regions will have.
Fascinating. Let me ask you, though. Then if you take a look at regulation in the US, it sounds like they're leaving it at the moment for the industry to self-regulate. I'm sure the other-- you're going to have some critique on the other side of maybe going too far in the other direction. But what do you see there?
Yes. Every region makes its own mistake when it looks at regulating something as broad as AI. And the US has a very laissez faire, hands-off approach to regulation in general, but particular with AI. Part of this reflects enormous power and influence of the tech industry, the amount of money they spend lobbying in DC, and even something like social media. It's very well documented, for at least the last 10 years, the harm that has, particularly on younger children.
Two years ago, they did draft the Kids Online Safety Act. That has gone nowhere. And that's terrible. So I think even for clear cut cases where AI and digital tech is doing harm, we've had a very difficult time drafting effective regulation. So I think it's hard to be hopeful that the US is going to be moving in a direction to regulate even pretty obvious harm cases.
Well, you have the two extremes. I'm sure somewhere in the middle is where things will get eventually. But let me ask you, though, if we could just maybe turn to AI in the markets. To say the promise of AI has been compelling would be the biggest understatement I think I've ever said. You take a look at the markets to see what's actually been going on.
Yes.
Are we in a bubble?
So there's three reasons to think we're in a bubble. One is valuations are stretched. The second is the degree of concentration. So for example, right now, the top 10% of companies represent 75% of market cap. That's only happened two times before, once in the late 1920s and second in the late 1990s. And both of those cases ended in tears. The third reason is the overall tone of hype about AI.
So those are reasons to think we're in a bubble. But there's two reasons to think this time is, in fact, different. One is the tech sector is producing lots of free cash flow, very impressive returns on vested capital, margins in the 20s. So that's very nice and very different from tech in the late 1990s. The second reason why is when our analysts visit companies, at least half of companies are offering sort of proactively examples to our analysts and to other people of what they're doing with AI to improve efficiency and productivity and how this is hitting the bottom line.
And this could be in terms of coding, writing marketing copy. It could be in terms of creative content. It could be video, music, a host of different ways that they're doing this. And this is very promising. And it will take time for this really to spread out, and diffuse, and have a major impact on overall productivity and earnings. But it is happening. And it's happening at a lot of companies. And it's quite possible that this is, in fact, being underappreciated in the market.
I wouldn't mind, Kevin, I know you've supplied us with a couple of charts, just to bring them up. The first one, taking a look at why people might think this feels like a bubble. When you compare it to past bubbles, at the top line, it certainly does look bubbly from one there. But I think if we move to the next chart, just to talk about what you're talking about is just the return that we're seeing on some of these is pretty compelling as well.
Yeah, if you look at the difference between the two lines, return on vested capital versus the cost of capital, that represents the economic value added by these companies. And it's enormous. And then if you looked at the S&P companies outside those seven, they're all struggling to have a return above their cost of capital.
So they're-- there's no doubt something special is happening not just with tech companies, but also companies that are being able to implement AI in ways that improve their efficiency, either helping earnings or reducing costs. And we are seeing lots of examples of that. But it's going to take a while for that to flow into returns. So at least for now, it's the infrastructure companies, cloud companies, platforms, semi-companies, which are really benefiting from the infrastructure period, the picks and shovels of the AI gold rush, if you will.
[AUDIO LOGO]
[MUSIC PLAYING]
[AUDIO LOGO]
As businesses push ahead with ways to develop artificial intelligence, governments are looking at ways to regulate its impact. The EU is amongst the first, recently approving a major set of regulations. But not everyone agrees this could be the best approach. Joining us now is Kevin Hebner, Global Investment Strategist at TD Epoch, to tell us what he sees when he looks at the industry. Kevin, it's always a pleasure to have you here. Can we just start off with, how effective do you think some of the proposed legislation coming out of the EU is going to be in regulating AI?
I don't think it's going to be very effective. There are some good aspects of the regulation they have. For example, in terms of AI-generated content, it's going to have to tell you that this is from AI. It's not something that a person created. Or if you're talking to a chat bot, it's going to have to tell you you're talking to a chat bot, not a real person. So those seem like good things. But overall, they're trying to regulate this very broad technology.
It'd be like regulating electricity, or the computer, or internet rather than regulating specific uses of it. And if you try to address every particular potential harm from something as broad as AI, I think the legislation is being so broad, covering so many activities, that it's going to be ineffective. Moreover, AI is developing so rapidly, we don't really know where it's going to be 2, years, 5 years, or 10 years from now. And I think that's a big issue. And every time you have regulation that's premature, overly broad, it ends up doing a lot more harm than good.
We can talk-- there's definitely some different tacks, I would say. We know in the States, they're going a little more laissez-faire so far, I will say, in terms of what's coming out. We'll get to that in a second. But if we take a look at that came out of some-- from an EU regulator, they talked about how the world should maybe look to the EU for a basic model of how to balance freedom and safety, not to focus on just the technology, but on risks likely to accompany its various uses.
Yes.
What do you hear when you hear that?
Yeah, so that is the perspective of the key architect of the EU legislation. So very much focused on risk, potential harms, not paying nearly enough attention to the benefits, the innovation, that comes with AI. And I think this is a big worry. When you look at the EU, they're regulating themselves to last place. They have very few tech companies. There's lots of very good AI people from Europe. But most of them now work in the United States. A couple smaller companies in Europe, but relatively little.
And I think given that these big technology innovations, they're why we're rich, they're why we have jobs, why we're relatively well paid, I think it's a big concern. And in particular, it's going to harm relatively younger people in the EU. And they're not going to have the opportunities that people in Canada, the United States, and other regions will have.
Fascinating. Let me ask you, though. Then if you take a look at regulation in the US, it sounds like they're leaving it at the moment for the industry to self-regulate. I'm sure the other-- you're going to have some critique on the other side of maybe going too far in the other direction. But what do you see there?
Yes. Every region makes its own mistake when it looks at regulating something as broad as AI. And the US has a very laissez faire, hands-off approach to regulation in general, but particular with AI. Part of this reflects enormous power and influence of the tech industry, the amount of money they spend lobbying in DC, and even something like social media. It's very well documented, for at least the last 10 years, the harm that has, particularly on younger children.
Two years ago, they did draft the Kids Online Safety Act. That has gone nowhere. And that's terrible. So I think even for clear cut cases where AI and digital tech is doing harm, we've had a very difficult time drafting effective regulation. So I think it's hard to be hopeful that the US is going to be moving in a direction to regulate even pretty obvious harm cases.
Well, you have the two extremes. I'm sure somewhere in the middle is where things will get eventually. But let me ask you, though, if we could just maybe turn to AI in the markets. To say the promise of AI has been compelling would be the biggest understatement I think I've ever said. You take a look at the markets to see what's actually been going on.
Yes.
Are we in a bubble?
So there's three reasons to think we're in a bubble. One is valuations are stretched. The second is the degree of concentration. So for example, right now, the top 10% of companies represent 75% of market cap. That's only happened two times before, once in the late 1920s and second in the late 1990s. And both of those cases ended in tears. The third reason is the overall tone of hype about AI.
So those are reasons to think we're in a bubble. But there's two reasons to think this time is, in fact, different. One is the tech sector is producing lots of free cash flow, very impressive returns on vested capital, margins in the 20s. So that's very nice and very different from tech in the late 1990s. The second reason why is when our analysts visit companies, at least half of companies are offering sort of proactively examples to our analysts and to other people of what they're doing with AI to improve efficiency and productivity and how this is hitting the bottom line.
And this could be in terms of coding, writing marketing copy. It could be in terms of creative content. It could be video, music, a host of different ways that they're doing this. And this is very promising. And it will take time for this really to spread out, and diffuse, and have a major impact on overall productivity and earnings. But it is happening. And it's happening at a lot of companies. And it's quite possible that this is, in fact, being underappreciated in the market.
I wouldn't mind, Kevin, I know you've supplied us with a couple of charts, just to bring them up. The first one, taking a look at why people might think this feels like a bubble. When you compare it to past bubbles, at the top line, it certainly does look bubbly from one there. But I think if we move to the next chart, just to talk about what you're talking about is just the return that we're seeing on some of these is pretty compelling as well.
Yeah, if you look at the difference between the two lines, return on vested capital versus the cost of capital, that represents the economic value added by these companies. And it's enormous. And then if you looked at the S&P companies outside those seven, they're all struggling to have a return above their cost of capital.
So they're-- there's no doubt something special is happening not just with tech companies, but also companies that are being able to implement AI in ways that improve their efficiency, either helping earnings or reducing costs. And we are seeing lots of examples of that. But it's going to take a while for that to flow into returns. So at least for now, it's the infrastructure companies, cloud companies, platforms, semi-companies, which are really benefiting from the infrastructure period, the picks and shovels of the AI gold rush, if you will.
[AUDIO LOGO]
[MUSIC PLAYING]