Artificial intelligence: a craze or a huge investment opportunity?

Hi, I’m Joss Murphy, research analyst at FundCalibre. Today I’ve been joined by Chris Ford, manager of the Sanlam Global Artificial Intelligence fund. Hi, Chris, how are you?

[00:13] Hi, Josh. Nice to see you.

Let’s kick it off then, Chris. I always find it surprising just how much AI is influencing so many parts of our lives at the moment. Can you tell us about a few areas? Maybe start with automation and reshoring?

[00:30] Yeah, so thanks Joss, it’s touching more and more parts of our lives, as you say. And this is something that we’ve been seeing throughout our experience running the fund now for five and a half, six years. And one of the areas which is most touched by AI right now, and very much in the Zeitgeist is, as you rightly say, the trend towards nearshoring, particularly in North America, but also in Western Europe – so, bringing supply chains back closer to consumers. And of course, there are costs associated
with that, and many of the facilities in North America and in Western Europe are going to be in high-cost areas, and so, finding ways in which those costs can be mitigated, particularly in the subsequent operation of a distribution facility, for example, is absolutely key.

And in many ways AI is all about trying to help us to do more with less. And so, AI’s benefits, as companies look to reshore/nearshore their supply chains, are absolutely manifold, and we’re seeing it deployed right the way across the supply chain from widget manufacturers all the way up to the OEMs [original equipment manufacturer] and of course, all those companies in between those people who are managing the supply chain, managing the flow of goods around the world.

So, for example, we’re finding it being deployed increasingly in the logistics supply chain by companies like GXO [GXO Logistics, Inc.], for example, who are a leading-edge provider of logistics facilities, to help companies allow goods to flow more efficiently through their warehouses. But also to allow companies, particularly in the consumer space, to be able to handle returns much more effectively, and doing that in a fully automated way. For example, if you bought a pair of Nike trainers and you don’t like them, you send them back for one reason or another. Then having a piece of artificially intelligent software that enables the receiving distribution centre to look at the goods as they’re returned, identify them for what they are, see whether they’re damaged in any way, and then basically send them off to the right place without any human having to actually engage in that process, is a huge benefit, not just in terms of money, but also time taken to get those goods back into circulation again and make them available to resell. So, just one example of how we’re seeing it manifest itself into the supply chain.

We’re also, of course, seeing those companies that sell the equipment into that process do extremely well. So, companies like Keyence [Keyence Corporation], for example, a Japanese leader in the world of machine vision, is really seeing very, very strong orders for its equipment. Cognex [Cognex Corporation], a competitor of Keyence in the US [is] also doing similarly.

We’re seeing great demand for companies that can provide the helping hand to those in the supply chain who perhaps don’t necessarily have all of the expertise internally themselves to be able to navigate their way into this world. So, companies like Hitachi [Ltd.], for example, with a lumada platform [a platform that provides solutions in multi-cloud environments] and their recent acquisition in the IT services space, are now in a position to be able to help companies find their way into the AI world and select which technologies are appropriate for them. And so, I could go on, as we rearchitect the global supply chains, AI is going to be absolutely everywhere.

And I suppose the last broader point I would make, is that this is a function of something which we see right the way across the economy. And that is that if you were to be building a facility or were to be starting a new business from scratch today with the benefit of a blank sheet of paper, why would you not design that business with AI absolutely at its heart? And, of course, you would. So, the question for many of these companies that all the companies is that how do you – without that benefit of a blank
sheet of paper – how do you bolt on AI as an afterthought onto existing business processes, to enable those benefits to be seen? And we’re seeing that happen faster and faster and faster.

No, that all certainly makes a lot of sense, Chris. But do you have an area that might surprise our listeners?

[04:27] I think it’s difficult to point to parts of the economy, which are completely untouched by AI at the moment, which I think is something which surprises me certainly. You know, the consistency and pervasiveness of AI within the economy, has been quite  extraordinary in its growth over the course of the last half decade.

Somewhere where we’ve done quite a lot of work more recently is in the agricultural markets where, you know, perhaps seen as a bit of a more sleepy part of the economy. But of course, actually if you look back over the last couple of hundred years, you see that agricultural markets have been some of those most extraordinarily disrupted by successive waves of technology. Firstly, obviously mechanisation in the 19th century and the late 18th century, and right up to the current day.

So, there’s nothing that suggests that agricultural markets, and the agricultural industry more broadly, are in any way less capable of being disrupted or indeed embracing new technologies than other markets. And we’re finding that some of the things that are being delivered into the market now by the agricultural equipment manufacturers, such as John Deere [Deere & Company, does business as John Deere] in particular, where we have a position in the portfolio, but also some of their competitors like Case New Holland [CNH Industrial N.V.] or AGCO [AGCO Corporation], enabling farmers to do more with less again: you know, use less water, use less pesticide, use less fertiliser – change the way in which farmers pay for services and in so doing, change their business model [and] change the margin structure inside of these businesses over time.

All of these are completely transformational, not just for the farmers, but also – particularly in why I care about it, of course – in an investment sense for the companies that are able to sell those services and products. And of course, as we see with John Deere, over time, as the margin moves up, as the services mix increases in the revenue line, the market will pay a higher multiple for that, we believe over time. And that’s what we’re beginning to see. So, I think agricultural markets are something which are worth
definitely keeping an eye on them.

Of course another big theme is climate change. What’s AI helping with there? I know Google has just launched Flood Hub. What’s that and [what else is out there]?

[06:38] So, Flood Hub is a piece of artificial intelligence software that Google developed to help identify, manage, and mitigate risk associated with sea level rise, effectively, but also other kinds of more localised flooding. And it’s probably worth reflecting on the fact that, for a very long period of time, actually meteorological information analysis has been in the vanguard of the delivery of artificial intelligence systems. We have the Met office in the UK, but also other weather services elsewhere in the world have been using
leading edge AI techniques for a long time now, using them to inform their predictions. So, there’s nothing new really particularly in what Google are doing, but what they’re doing is making that leading edge AI available to a broader group.

AI can help in a number of ways with climate change. One of the ways that you’ve already identified is how you highlight the impact of climate change and forecast what that looks like forward, and then mitigate for it.

But of course, there are other things that AI is able to do in particular, again, coming back to this point of doing more with less; one of the things that Alphabet [Inc.] has been doing for a long time now has been using its own AI to manage its energy consumption inside its data centres and Alphabet’s data centres some of the most efficient anywhere in the world, in respect of their energy consumption. And that, of
course, is a huge hot potato at the moment, as people now begin to focus on the huge energy requirements which are demanded by data centres around the world.

And I think the most recent ChatGPT [a chatbot computer programme] conversation has really shone a light, very brightly on that, where leading edge, artificially intelligent workloads are highly consumptive of high-powered computing capacity. And there’s a cost to that, and that cost comes in the form of energy consumption. So, how we call those facilities, how we power those facilities, is absolutely key. And AI has a huge role to play in helping us to understand how that works. And of course, that then begins to feed out into the broader electricity supply chain more broadly. So, the management of distribution networks, the management of power generation facilities, all of that is
a part of the economy, which AI has a huge amount to offer. And, of course, that then begins to map back into the meteorological conversation that we were having earlier. If you can better forecast the weather, if you can better understand what wind and solar generation rates are likely to look like, then you can better manage your energy generation and configure the grid accordingly.

So, you know, there are nested levels of benefits that AI can bring right the way across this climate change discussion from managing energy use, managing energy generation, and then figuring out what we could do to mitigate the impacts of climate change where we need to.

No, that’s certainly hard to disagree with Chris. Looking back at AI’s development over the past decade, we’ve gone from Google’s brain learning how to find cat videos in 2012 to Alexa teaching motor skills through robotic hands and winning chess games. What’s next for AI? What’s exciting you today?

[09:48] I think, well, you mentioned chess. You know, there’s always the ability to trace the history of AI through its ability to overcome games of succeeding complexity. And of course, you know, chess was 30 years ago now, almost – you know, we forget that these things are … things have already gone on. We now have platforms that are capable of playing Go, but also capable of playing StarCraft and other massively multi-player online games. So, hugely complex game spaces where AI is capable of engaging. I think that absolutely continues. And for game players, that will be great, it’ll just mean a richer AI
experience inside games.

I think some of the things that we’re seeing now with NLP, natural language processing, of which ChatGPT is the most recent example to have caught the eye, are really important. And, you know, we are perhaps a little less surprised and taken aback by the capabilities of ChatGPT than others because we regard this really to be just a natural progression that we’ve been able to observe occurring over the course of the last 36 years or so, in the efficacy of natural language processing. But ChatGPT is extremely good, and it is beginning to cross a rubicon I think now, to the point whereby the benefits that ChatGPT can deliver are very apparent and anybody who’s had the opportunity to engage with that platform – and I heartily suggest you do! – will have seen what it’s capable of achieving.

That technology we think will be coming into Microsoft Office products over the course of the next couple of years. And that will be really interesting because at that point, when you have that kind of level of natural language processing embedded inside the office suite that of course we all use every day in the course of our work, you know, that’s now beginning to deliver fully functional artificial intelligence systems into a front office context, in a way that we will all engage with on a daily basis.

And that’s a really exciting proposition. So, I think that’s coming. I think we’re going to hear more from Alphabet as well in respect to natural language processing and there’ll be a real kind of tug-of-war between Microsoft and Open AI on the one hand and Alphabet and DeepMind [British AI subsidiary of Alphabet Inc.] on the other for supremacy here, and that’s just what’s been going on for the last few years.

I think the other thing I would point out would be what’s going on in drug discovery, which we think is really, really interesting and very important. And so, a couple of years ago now, Alphabet developed a platform called AlphaFold [an AI program developed by DeepMind] which was a platform developed to address the problem of how to predict and model the way in which proteins fold inside the human body initially, but potentially in other organisms as well.

And the platform – without going into huge detail here – has now been carried forward to the point that Alphabet actually established its own drug discovery or operation called Isomorphic Labs under the auspices of DeepMind last year – sorry, 2021. And Isomorphic Labs is the platform that Alphabet hopes to use AI to significantly increase the efficiency of drug discovery over the course of the next decade. And that was an incredibly important, exciting proposition for all of us, frankly. And I think Alphabet has behaved really extremely well as a good corporate citizen, in respect of making much of that AlphaFold intellectual property available for others to scrutinise, so that it can be understood what it’s looking to do.

Chris, I just want to say thank you so much for giving up your time today.

[13:23] Not at all.

And if you’d like to find out more information about the Sanlam Global Artificial Intelligence fund, please visit FundCalibre.com.