This is not a piece about banal AI-written investment content, it is an exploration of why a reliance on AI kills the benefits of writing for both reader and author.

When I realise that I am reading a piece of writing that was AI-generated my heart sinks a little. In part this is because of the frustrating style it adopts with its incessant use of antithetical parallelism – “this is not just X, it is Y”. Yet there is something more important than this tedious rhetorical-tic: AI investment content strips all of the value and purpose out of the act of writing and reading.

Most investment writing (apart from generic, turn the handle market commentaries) should be thought of as a process which benefits both reader and author. The reader gets to understand the voice and views of the author, while the author can write as a way of shaping and developing their own thought processes.

If investment writing is just the result of a few prompts, it lacks any original thought and fails to represent the author’s voice. Rather it is simply the output of a predictive system that has no stake in what it generates.

For the author, high quality writing should never be merely an exercise in creating some form of content. Of course, writers are trying to communicate something, but the path to arriving at that communication is as important as the message itself. The process of writing involves exploration, reflection and refinement – all of this is lost if we simply rely on AI.

This loss is most pernicious for those with less experience or without true domain expertise. The process of writing to communicate is also an invaluable form of learning. When the answer is simply to ask an AI model to assume this responsibility for us, not only do we sacrifice the opportunity for slow, deliberative learning, but we also do not have the means or knowledge to question whatever answer AI provides. Writing is about so much more than the output.

Even if we do think only in terms of the end product, the loss from AI-reliance is pronounced. Readers know nothing about the person that (they think) might have written something. They have no sense of their voice, nor can they even provide feedback or challenge to the author, because it is not the author’s work.

Successful investment writing comes largely from a relationship of trust between reader and author that builds over time. As readers, we pay attention to the words and opinions of someone we might have followed for years, and we make judgements about the quality of views offered based on that trust. In a world where AI investment writing dominates, there is no trust, just an absence of it.

While the process of writing informs and enlightens both reader and author, writing can also act as a useful signal of expertise, effort and care. The friction that exists because good investment writing is difficult and complex evaporates when AI can churn out passable flannel in a few seconds.

It is possible to argue that the genuine, higher quality writing will shine through amidst the barrage of generic babble, but it will become increasingly difficult to find amongst the maelstrom of AI generated pieces.

AI writing will also inevitably get better and seem more authentic, so it will become harder to spot than it is now. This will not be a positive development.  As the quality of AI writing improves, the loss from our increasing reliance on it will become more pronounced. As more people turn to it, more of us will lose the benefits that stem from authentic, human writing.

There are, of course, aspects of the process of writing where AI can be incredibly helpful such as researching, fact checking (sometimes) and proof reading. It is also useful in producing the generic work that I imagine no human enjoys having to create, but in writing, and in general, we really need to start being very clear about what AI is good for and where its presence comes with a significant cost.

Investment writing should have a clear purpose and one that is more than simply attempting to communicate a certain message as quickly as possible. For many authors it will be a way of thinking through messy, subjective and complicated problems. While readers want to hear the trusted voice of a writer and understand what they are thinking and why.

All of the meaningful elements of investment writing are human, and most of these have very little to do with the simple production of content. As we increasingly rely on AI, we risk losing the manifold benefits that come from the process of writing and reading.

—

---

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. Not investment advice.

Although equity markets have thus far proved remarkably resilient in the face of geopolitical and economic risks, it is hard not to argue that we are in an environment of heightened uncertainty. Dispersion within equity markets has been extreme, asset prices are moving dramatically on a single tweet, and there is a daily torrent of erratic news flow. This is clearly a difficult backdrop for investors. Perhaps its most worrisome feature is the temptation to do two things we really shouldn’t.

What are these two things?

• Timing market movements.

• Concentrating our portfolios.

These are both usually bad ideas, but they are particularly pernicious in febrile financial markets. Unfortunately, our desire to undertake such actions seems to be strongest at the worst possible moments.

Time to Time

It seems odd that a reaction to rising market uncertainty is an increasing desire to predict how the market will move, but this is inescapably the case. We have an inherent discomfort with ambiguity, and by investing as if we know the future it can give us a sense of control. We can’t bear to sit and let markets happen to us.

To add to this, we are also loss averse. When people talk of market uncertainty what they really mean is that investors are worried that the chances of losing money are increasing. This is why most market timing behaviour is an attempt to get out of risky assets before a major fall, and then get back in at just the right time.

Trying to invest in such a fashion is a sure route to bad outcomes. Not only is the chance of us getting the initial timing correct incredibly low, but even if we are right we then need to keep making similar decisions. Getting in, and back out again. And what about next time? Do we try to repeat the trick every time we feel that markets are at a precarious point?

We have an incredibly strong instinct to act and to do ‘stuff’ – don’t just sit there, markets are moving! For investors that often means trying to make predictions about what happens next. While the urge to do this can feel irresistible, it flies in the face of the evidence about what is likely to work for us over time.

Lacking in Concentration

It is not just the temptation to trade and predict that arises during periods of heightened volatility and unpredictability – we are also likely to become more concentrated in our portfolios. If we see wide levels of dispersion across stocks, styles, asset classes or sectors, our instinct is to think: “Why am I owning these laggards? Why shouldn’t I focus on the winners?”

This is another form of market prediction, but in a different guise. When we see major performance dispersion in areas of the market (such as energy and software recently), and the compelling stories that are used to explain it, we can’t help but extrapolate. If this trend is going to go on forever, then all we need to do is focus on the right parts of the market.

Losing sight of the principles underpinning prudent diversification is a dangerous game, and can be extremely punitive in choppy markets where dispersion is wide. The cost of not holding certain assets or areas of the market can be heavy if we end up on the wrong side, which at some point we inevitably will.

The idea that diversification is a free lunch has always been a naive one, because it ignores the behavioural element. Being diversified means owning things that are performing poorly and that, in certain environments, we should expect to perform poorly. It is incredibly difficult for investors to embrace this idea – we just want to hold the good things that have been working.

Although I have framed market timing and concentration as distinct activities, they are – in essence – both a form of overconfidence in our ability to predict the future. It’s not that investors are gaining confidence as markets become ever more tricky to anticipate; it is rather that the sense of anxiety stemming from rising volatility makes us behave in ways that appear as if our confidence is on the rise.

Attempting to engage in more difficult activities as a reaction to a feeling that conditions are becoming increasingly challenging is an odd and costly path, but a very human one.

—

The sense that markets are unusually uncertain is the time when behavioural discipline matters most, but also when we are highly likely to abandon it. Market timing and portfolio concentration feel right, but really they are just the easy options that make us feel good, while being very unlikely to work.

—

---

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. Not investment advice.