Do I risk being labelled an old-fashioned, out-of-touch Luddite if I don’t whole-heartedly buy into the AI revolution? Perhaps I am simply long in the tooth or sufficiently seasoned enough to tolerate the inevitable criticism that comes with suggesting the industry may be getting complacent. As Edward Luce recently wrote in the FT, for AI cheerleaders, “safety is woke”; so I open myself up to such a label.
That said, I’m not entirely sure the AI enthusiasm in corporate America and on Wall Street is universally shared. Step outside the industry bubble and there’s a noticeable degree of unease about where all this leads. Whether justified or not, that disconnect alone should give us pause for thought. Markets tend to run into trouble when confidence becomes one-sided.
Back in the day (indeed, in a millennium far, far away), one of my jobs as a fresh-faced graduate was to attend the fax machine at midday, awaiting the daily ‘research note’, a centrally aggregated summary of broker output, that took a morning to be composed and distributed to all the regional offices.
A few years later, now firmly settled in London, a rite of passage for new team members was opening the post and compiling a list of all the broker notes delivered that morning. And of course there were the routine broker calls, a thing of the past in this post-MIFID world.
It seems very quaint now. But my point is that, in that analogue environment, you had very tangible, daily contacts with the market vibe. Through sheer exposure, you absorbed context and developed judgement. It was time‑consuming, but it was effective. It was a great way to learn. And, counterintuitively, you learned fast. I’m certainly not suggesting that way was better, but the journey was the fun part.
From immersion to instant answers
Today, time itself appears too precious; immediacy commands a premium. There are now far more efficient ways to process information. When a company reports results, an AI agent can instantly produce an executive summary, generate charts and tables, and even curate a set of questions to prepare for management meetings. More advanced tools can compile initiation notes that combine company disclosures with sell‑side analysis. It is fast, comprehensive, and frictionless.
This gives me the heebie-jeebies. AI is undoubtedly a powerful tool, and I use it myself – I used it to help edit this article. But it’s not a substitute for independent analysis. I believe there’s a danger that we corrode what’s truly great about active fund management. Of course, the efficiency afforded by AI is valuable. But we don’t want to lose the gift of the time, space – and ultimately the capacity – to think.
I’m still a relative neanderthal. I still print off company results. It may not be the most environmentally friendly approach, but it allows for deliberate reading. I can underline. I can scribble notes and comments. I engage with it. This is my way of assimilating the information. The best questions you can ask are the ones that you’ve composed yourself – the act of constructing them is itself integral to critical thinking and therefore to any investment process. No investment decision is infallible, but by doing this work, by formulating an argument through careful thought and discussion, it’s my way of having a fighting chance of spotting when and where a thesis may unravel.
Is AI a passive tool?
A wall of studies shows that the act of writing something down makes you much more likely to remember it. You listen, you determine the meaning, you summarise, and you write it down. This is tremendously active. Today, AI can record an entire meeting and then do all those other steps that help us recall. If we’re not careful, it will make us into passive listeners.
AI is an incredible tool, but it is unlike other tools. AI is inherently passive, unlike other new technologies that we’ve developed over the centuries. You give it a task, and away it goes and does it. That can free you up to do something else useful in the background. Or it could allow you to kick your feet up. When it comes back with an answer, you must probe it. But you haven’t done all the steps that would have helped you take an informed opinion of your own. In short, you’re less able to push back on it because it’s done all the work. And this gets exponentially worse the less you know about a subject.
We need to find the balance between harnessing AI for efficiency's sake in the short term and keeping our minds strong and supple in the long term. Skilful use of AI helps to achieve that balance.
AI’s hidden biases matter
We don’t want to be involved in a race to the bottom. If market participants converge on the same AI tools, the risk is homogenisation of insight.
We should be wary of treating AI tools as the sine qua non of active fund management. Relentless efficiency risks becoming a misguided Trojan horse, positioning active managers in direct competition with algorithms and passive strategies, and I don’t think we win that game.
On the other hand, going against the grain and doing the donkey work can be a point of difference. It does take time and effort. But what is active fund management if it isn’t taking the time to think, to argue, to discuss, to formulate? This isn’t a weakness to obscure, but a discipline to champion. If you can meld this with the valuable tools that AI afford, there is a winning combination for active fund managers.
And, may I humbly suggest, it may also be a more intellectually rewarding, and indeed fulfilling way of going about our business.
