Where do the opportunities lie?
AUDREY RAJ:
When we think of AI, there are sectors that are heavily invested, such as industrials and healthcare. In many cases it’s solving problems, analysing masses of data. Where are you seeing investment opportunities?
LAURENT LEQUEU:
I share the view that healthcare is a sector experiencing big developments. I also think that finance is an area to watch, and Big Data can be a game-changer for the industry, even if it is only in the early stages for the financial industry.
CLIVE McDONNELL:
My view’s a bit different. I don’t think of using Big Data to generate viewing or shopping suggestions as true AI. Rather, I see it as using algorithms to enhance the user experience. Whether it’s shopping or another online activity, the data that’s created is enabling a continuous improvement in the algorithm, which can enhance the online experience and make it more rewarding, but I don’t see that as AI per se.
There are some areas where purer AI plays are available, such as medicine, where scientists are beginning to programme computers to learn independently. Whether it’s for cell therapy, cancer, gene therapy or other things, that’s very much at the early stage.
I think the distinction between AI and automation has been blurred over the past few years, but I think they are two very distinct things and the investment opportunity does exist in the automation space, which is more robotics, as opposed to AI. That is a good thing because there are more opportunities in listed companies in robotics than in AI.
BRYAN GOH:
I actually struggle to put my finger on how you take advantage of AI. Look at technology, for example. You can go back to the 1960s and 1970s and call that a PC revolution, but eventually it became so ubiquitous that everyone used it, so it loses its premium. This is what is happening now. Look at what happened with the S&P and its redefinitions. Amazon is now no longer a tech company, it’s consumer discretionary.
Google and Facebook are no longer tech companies, they’re telecommunications and media. So what happens when you’ve got this technology that is so successful, so ubiquitous? Where’s the premium in owning the pure tech? I struggle with that. I don’t have the answer to be honest.
AUDREY RAJ:
Is that why you might co-own a robot?
BRYAN GOH:
Yes, or you can own one yourself. What’s the meaning of AI though? I still don’t know what it really means. Right now, we have algorithms that are capable of very core-oriented tasks. Is that AI? I don’t know, I don’t know how you classify that.
MARTIN FORD:
I think that’s right. There’s a very common saying in the field that once AI becomes very practical and it’s used all around us, we don’t call it AI anymore. That’s what’s happening here. The algorithms we’re referring to are very often deep learning algorithms – they’re neural networks. So they definitely all rose out of AI research, but once they become practical we don’t refer to them in that way.
The same is true of expert systems. Aeroplanes that can fly themselves – those are expert systems. At one point in the past, that was considered AI too, but we don’t really refer to it in that sense anymore.
But you’re right. There is a distinction between the cutting edge of AI and machine learning, and what you see in traditional automation.
SEBASTIEN GENTIZON:
Diversification seems important at this stage to try and capture the winners. But after we enhance those industries, whether it’s IT or banking or medicine, sometimes the winner takes it all. That’s certainly what has happened with some of the latest tech revolutions. So when you look at the smaller players today, what do you think will allow them to compete with larger players? Or will they become acquisition targets instead? Those are some of the key questions you should ask yourself when you invest in those fast-changing industries.
LAURENT LEQUEU:
I think the risk is that governments will want to regulate these big players. We are back at the same story from many decades ago when there was only one company, Standard Oil, which controlled the oil supply. The same is happening now when it comes to the data that powers AI and machine learning. This may be underestimated by the market, but we could see further regulation coming for big players such as Alphabet or Amazon.
MARTIN FORD:
There are good reasons to believe that big companies such as Alphabet, Facebook, Tencent and Baidu are going to continue to be dominant. In fact, they seem likely to buy any start-up that really demonstrates progress in that area. DeepMind, the British AI company acquired by Alphabet, would be one high-profile example.
But everyone overlooks the fact that data is vertical. If you have a health insurance company that owns lots and lots of data about patients, Google doesn’t have that data. It doesn’t belong to Google. Now, Google may supply that technology in its cloud server, but the data is owned by the health insurance company and they’re the ones that are going to own that value, they’re going to create it. So it’s really a mistake to look at AI and say it’s all just Google, Facebook and Tencent, because the value is going to be distributed vertically. It’s really important to understand that there are lots of other places where value is going to be created, even if a few companies are actually creating the core technology that enables that.
AUDREY RAJ:
What about start-up companies? Are they a good place to invest as well?
DANIEL FERMON:
It’s difficult to invest directly in start-ups, as it’s a private equity business. We looked at it, but we didn’t find a good way to invest for us. It’s definitely a growing area. But so far it’s limited. All the start-ups are acquired by media companies. Over the past seven years, more than 250 companies have been acquired by Google, Baidu and the like.
MARTIN FORD:
We invest in companies such as Google and Baidu, and they have a tremendous amount of expertise in finding these start-ups, finding the best ones to buy. They’re much better at it than we would be. At least for an ETF investor, making sure you have exposure to these companies is a pretty effective way to do it.
DANIEL FERMON:
That’s true. In 2014, DeepMind wanted to IPO, but Google came and said: ‘How much?’ So at the end of the day, it’s difficult to invest in that space and to be ahead of these giants.
