'AI will change our world in more ways than we can imagine'

(Image credit: Getty Images)

Andrew Van Sickle: A recurrent theme this year, and the subject of one of your key research papers, was drawing parallels between the dotcom bubble and this AI boom. What’s your take? Is this a case of rational exuberance?

Rob Arnott: It is a case of rational exuberance, with somewhat irrational pricing. To be clear, AI is the real deal. There is no ambiguity about that. It will change our world in more ways than we can imagine.

It was the same story in 2000. The internet was going to change everything: how we buy and sell goods and services; how we communicate with family and friends – or clients; how we get our news; how we do our research. And all of that happened. It’s just that the internet wasn’t embraced as rapidly as its cheerleaders expected, and I think the same will be true with AI.

MoneyWeek

Subscribe to MoneyWeek today and get your first six magazine issues absolutely FREE

Get 6 issues free

https://cdn.mos.cms.futurecdn.net/flexiimages/mw70aro6gl1676370748.jpg

Sign up to Money Morning

Don't miss the latest investment and personal finances news, market analysis, plus money-saving tips with our free twice-daily newsletter

A vital issue the narrative in 2000 missed was that the biggest stocks at that stage weren’t necessarily going to be the leaders in 2010, let alone 2020. I recall Cisco Systems’s chairman in 2000, John Chambers, saying he didn’t see why Cisco couldn’t grow sales by an annual 40% for years to come. At that pace you grow sixfold in five years. In fact, it grew sixfold in 25 years. Still impressive, but not 40%. As a result, investors holding Cisco since March 2000 are slightly in the red.

Of the top-ten companies by market capitalisation in 2000, only Microsoft remains on the list today. The other nine have all been displaced. Intel and Nokia were on the list, for instance. Other firms became better at producing chips and phones. People talk about companies having moats (competitive advantages that stop rivals from encroaching on their territory), but moats don’t last forever.

AI is already displacing current market leaders

There has already been some of this sort of displacement with AI. Google’s core business model is making money from sponsored links and pop-up advertisements. People discovered that ChatGPT or Perplexity can be used as a search engine, and you don’t get annoying pop-ups and sponsored websites; no need to scroll through a dozen sponsored websites to get to something useful. Moreover, OpenAI itself was disrupted a year ago by DeepSeek.

So this is a rational bubble in that AI is very real. All the market leaders have a shot at being dominant players for years. But I also recognise that some of them will be displaced.

Andrew Van Sickle: Could you give an example from your experience that makes you think AI is the real deal, as you say?

Rob Arnott: The latest iteration of ChatGPT is extremely impressive. When we write a research paper, we hand it over to AI and ask it to write a summary and critique it. We also get it to tell us if we’ve missed any references or citations that we ought to have considered.

We ran a paper we recently finished through ChatGPT and got an elegant synopsis that was better than anything the authors could have come up with. We also got a thorough evaluation of what was good about the paper and what wasn’t so good, along with a list of citations we’d missed. Four of them made us think: “Oh, my goodness, we should have thought of that one.” Four were borderline, and two didn’t exist.

The upshot is that AI can’t think like a person, but it’s excellent at scouring data and comparing it with information it has already absorbed. It can sift data sets of billions and billions of data points and then it will start to appear close to being human. If there are fewer data points, though, it will struggle badly.

For example, when it comes to researching long-term stock market behaviour there are only thousands or millions of data points, depending what we’re looking at, so AI can’t come close to the efficacy of human judgement supplemented by ordinary statistical tools. It can’t predict things like whether there will be a market downturn or recession. In short, AI is a quantity, not a quality game.

Andrew Van Sickle: Given the recent emphasis on technology growth stocks, where do you, as a value guru, see bargains?

Rob Arnott: The spread between growth and value is wide everywhere; globally speaking, it is near record territory. In the US, it is close to the level seen at the peak of the dotcom boom and the summer of 2020.

I would say don’t abandon growth, but don’t overweight it. Betting against a bubble is dangerous because bubbles can last longer and go further than you can imagine. So to the extent you have value in your portfolio, put more money into that.

We at Research Affiliates are always keen to find value and reduce concentration risk. We came up with the Research Affiliates Fundamental index (RAFI) series in 2005. These indexes are skewed towards value. The companies are weighted by their size and hence their economic importance – using measures such as sales, cash flow, book value and dividends – rather than share price. This approach has done much better than market-capitalisation-weighted value indexes, proving superior in three out of four years.

So investors might like to research the RAFI-based exchange-traded funds (ETFs) from Invesco that focus on value. There is the FTSE RAFI All World 3000 Ucits ETF (LSE: PSRW) for a global play; the FTSE RAFI US 1000 Ucits ETF (LSE: PRUS) for US value; and its European counterpart the FTSE RAFI Europe Ucits ETF (LSE: PSRE). The FTSE RAFI UK 100 Ucits ETF (LSE: PSRU) is a British version.

Andrew Van Sickle: Let’s move on to gold. It appears to be pausing for breath. Is the medium-term outlook still bullish, do you think?

Rob Arnott: The biggest factor that’s bullish for gold is the developed world’s addiction to debt-financed government spending. The fear is that fiat currencies will create their own financial crisis that could be bigger than the global financial crisis of 2009; I think that’s reasonably likely.

I don’t see it as an imminent threat. But just as Greece had to face the music at some point when its debt started to soar towards 200% of GDP, the US isn’t immune to that dynamic. We may be the biggest economy in the world, but that doesn’t mean we can endlessly spend money we don’t have. A crisis could arrive in ten to 15 years’ time.

It would require a severe bout of inflation to reduce the real value of the debt, and the consequence would be that the gold bugs of today would not regret their purchases even at today’s prices. That said, I am not a gold bull for 2026, given the recent surge to new records and the froth in the market.

Andrew Van Sickle: What could the result be of Donald Trump’s apparent desire to influence the level of interest rates set by the US Federal Reserve?

Rob Arnott: The upshot may be that he will keep rates 1% or 2% below where they ought to be, resulting in free money, which I define as rates below inflation: negative real rates. We had a dozen years of free money in the US, which was a key reason for slow growth.

The trouble is that stimulus doesn’t stimulate. My colleague Chris Brightman has written a great deal on stimulus failing to generate sustainable, healthy growth; you can find some of the papers on our website. Monetary stimulus leads to asset bubbles, exacerbating wealth inequality. It pushes money into the hands of people who are already well off, and they spend it, boosting corporate profits and stimulating the stock market further.

Part of the overall stimulus in recent years has been fiscal, too. That involves the government either taxing or borrowing money out of the private sector and then spending it, so it can hardly stimulate private enterprise.

Andrew Van Sickle: What is the outlook for US inflation? Will the tariffs boost it?

Rob Arnott: What could make inflation jump again is massive overspending, which we’re still doing, so that is a dangerous game. I wouldn’t view tariffs as particularly inflationary, however. I remember many analysts panicking early last year about a deep recession and hyperinflation.

But think it through. If the tariff rate settles at an average of 15%, it’s no different from a 15% VAT levy – applied in a limited way, solely to imports. That’s hardly catastrophically inflationary. It’s a tax. It boosts costs on taxed items. But if our imports in the US are worth 11% of GDP, and if we have an average tax of 15%, then it’s going to raise 1.6% of GDP in tax revenues. And that’s exactly what happened.

So who pays it? The suppliers swallow part of it to remain competitive in the US market; the supply chain takes some of the strain. The consumer has to pay the rest, perhaps half, or 0.8% of GDP – as a one-off. As a libertarian, I think the correct tariff rate is 0%. But the tariffs aren’t going to send inflation rocketing.

Trump’s tariffs games are essentially him implementing a strategy from Sun Tzu’s The Art of War: scare your adversaries and you can win without firing a shot. So, he threatens 100% tariffs and opts for far less. He’s lived by that book for his entire life.

Andrew Van Sickle: Last time we talked, you said you thought China had taken a statist turn under Xi Jinping. He has recently said he wants to make bolstering the private sector more of a priority. Has the regime realised that it will need to be more free-market to get China out of this debt deflation?

Rob Arnott: The trouble is that Xi Jinping is an unreconstructed Maoist, and so paying lip service to the private sector is about as far as he is likely to go. I still think China will get old before it gets rich.

This article was first published in MoneyWeek's magazine. Enjoy exclusive early access to news, opinion and analysis from our team of financial experts with a MoneyWeek subscription.

Explore More