As an investment theme AI has it all: the potential for transformational societal and economic change, huge capital investment and the prospect of dramatic corporate winners and losers. It would be hard to design a better setup for investors to lose a lot of money. 

I have no confident view on quite how profound the continuing developments in AI will be for our lives – but then I don’t think anybody really can. The problem is that many investors will be tempted to behave as if they do. 

All investment themes are dangerous, but a narrative as powerful as the one that currently surrounds AI requires a particular level of caution. When investors anticipate (or worry about) extreme change we find it hard to escape the feeling that our portfolios, and even our entire investment approach, needs a rethink. 

If there are to be winners and losers, then surely we must act to best exploit the opportunities and mitigate the risks? This notion sounds eminently sensible, but really isn’t. 

When we adapt our investment strategy in reaction to potentially dramatic thematic market shifts – such as AI – what we really mean is that we want to be more concentrated in our portfolio. Concentrated by idea, country, sector, stock – focused on all things seemingly related to the prevailing narrative.

Becoming more concentrated in our investment approach is simply a way of expressing that we are increasingly confident about the future. The more conviction we have in our predictions, the less diversified we need to be.

Unfortunately, taking a more concentrated approach at a time of (possibly) seismic shifts is precisely the wrong thing to be doing.

If AI is to lead to genuinely consequential change our preference should be to be more diversified not less.  Nobody can say with any confidence how it will play out, so why would our portfolio activity suggest that we can?

When investors talk about AI beneficiaries and focusing their portfolios on such areas, it is worth considering the critical questions that need to be considered:

• What will be the pace and scale of development in AI technology from here?
• What will be the economic and societal impact?
• How will it impact corporate profitability across sectors?
• To what extent are AI developments already reflected in stock prices?

These questions are just the start of what is a staggeringly complex topic. We should be ensuring that our portfolios reflect the sheer level of uncertainty that exists around AI, not make investment decisions that imply that the outcomes are self-evident.

There are, of course, a range of AI-associated stocks that have already benefitted grandly from the theme’s emergence, but whether these are anything more than first order momentum trades is fiendishly difficult to decipher. 

The temptation to adopt an increasingly concentrated investment strategy in order to best capture the impact of AI exposes us to two potentially disastrous risks. First is that the consequences of AI are less significant than anticipated and ‘AI related’ stocks are materially overvalued. Second is that AI is as transformative as many anticipate, but the companies that benefit are very different to current market expectations. 

It is always important to remember that being right about some economic or technological development does not mean that we will make money from it. The internet did change the world, and China did rise to become a global economic superpower.  Knowing both of these things in advance would not have been sufficient to deliver good investment performance. In fact, knowing these things in advance would probably have increased the risk of very poor returns.

Despite constant lessons from history of the dangers of making concentrated bets on seemingly inevitable themes, the lure of repeating such mistakes will prove irresistible to many. The combination of strong performance, and compelling stories will draw us in. 

There is, however, no need to have an ‘in or out’ view on AI. Being diversified allows us to benefit from holding the areas of the market that benefit most significantly from the progress of AI, while providing some protection if its impact is underwhelming relative to expectations, or it has economic or corporate impacts that we did not foresee.

The most powerful investment themes can often make us feel as if things are becoming more certain just as they are becoming more unpredictable and risky. Our best guard against the risks that might stem from unknowable change is to remain humble and diversify. The more AI hype grips markets, however, the more likely we are to do just the opposite.

—

My first book has been published. The Intelligent Fund Investor explores the beliefs and behaviours that lead investors astray, and shows how we can make better decisions. You can get a copy here (UK) or here (US).

All opinions are my own, not that of my employer or anybody else. I am often wrong, and my future self will disagree with my present self at some point.