AI – opportunity or threat?

CR: Hello, welcome to Silicon Valley Byte Size, an update on the tech sector from the Allianz Technology Trust. I'm Cherry Reynard. Today we're going to be talking about artificial intelligence. The rapid adoption of ChatGPT has brought it to investors' attention and share price performance for related companies has been very strong since the start of the year. However, recent reporting from the tech giant suggests it may take considerable investment to realise its potential. So, with me to take a closer look is Mike Seidenberg, manager of the Trust. Welcome, Mike.

MS: Good to see you again.

CR: Can we start with a look at how AI works and the various different types?

MS: Yeah, absolutely. I mean, there's so much that's been written, I'm sure, even for myself, sometimes my head spins. So, I think the right way to break it down is to think about where we are today, where we're going, and where we hope to be. You know, most of the AI today are reactive machines. These are machines where, you know - you've probably seen them whether you interact with a particular consumer brand or a business brand; and, they're asking you if you need help and, if you deviate at all, you usually just say, “agent”, right? Because you want to talk to a person. And, by the way, these machines are helpful, you know. I had a really good interaction the other day where I needed to make a return because I bought vanilla paste versus vanilla liquid; and on Amazon, it’s a non-returnable item. But knowing that on Amazon, you can return things, you just have to kind of go a layer deeper, and I interacted with a chatbot. And it was a really good experience. I finally figured out it wasn't an agent, but it took me a while and obviously, I got my refund, which is great because I felt great as a consumer. But I think that where the world is going to, is towards limited memory artificial intelligence. Which is why people are so excited about ChatGPT. Now, with the definition of limited memory, that tells you that there's only a certain amount of characters that this particular technology can handle today, and that continues to grow. As you see people write about ChatGPT and others; whether it's images, or words, etc, etc. And those will grow. Where we want to go to eventually, is more a kind of theory of the mind, right? Like, how do we as human beings interpret things? I read a great example of Google Maps. Whether you miss your turn or not, it's very rudimentary with respect to just rerouting you, but you know, really in a perfect idea it's to create machines that interact more effectively with humans and take into consideration our state of being. And we're not there yet, today. I think we'll get there. And obviously, that's a little bit down the road. But I think that what you're seeing, broadly speaking, is just use cases that are really innovative for companies primarily, because, you know, at the end of the day; and I say this a lot about technology, technology tends to be a way that companies create some type of either, you know, sell more things, happier customers, take costs out, right, it's kind of those three things that I always think about technology solving challenges. So yeah, artificial intelligence is really top of mind. And obviously, where the future goes, who knows? But that's kind of where we are today, in my opinion.

CR: And I mean, you mentioned those use cases. Are there certain sectors or areas where it has particular application?

MS: Yeah, great question. You know, images are something that a lot of people are talking about - specifically if you use a medical use case. Whether it's looking at lots and lots of images of a particular PET scan or a CAT scan and looking for something that no matter how good you are as a radiologist, using technology to basically augment your decision in being able to identify something that maybe the human eye really can't identify; and artificial intelligence allows that. And obviously, if we extrapolate that, the earlier we might find something in a particular individual, the better the outcome probably is. So, I think that's a good use case I think you’re looking at. If you think about another side of images is just making sure - on the advertising side - that the correct image is being served to the customer, right? Or to the individual. So really allowing the personalization experience. And I think you're going to see really clever things done by the companies that market to the folks like me, and you. And if I think about oil and gas, I think about interpreting log data, coming off seismic - some of the equipment they use out in the field. Just making better decisions as to potentially, where might you want to put a water injector in an oil field to increase production. Stuff like that, it really runs the gamut. And I think the reason why you're hearing me talk about various verticals, is really this notion that the interest level is across all businesses. It's not in one particular business, because I think in general, businesses are seeing applicability; and in use cases or test cases that really kind of open the horizon for their business, with respect to using it.

CR: Yeah, obviously, that's been part of the excitement since the start of the year, that there are so many applications for it. But Microsoft, possibly Meta as well, slightly spooked markets in their results statement when they were saying there's a lot to gain out of AI. But, you know, there's going to be a lot of costs involved as well. I mean, where are we on that sort of cycle? Is there still considerable investment needed to realise its potential?

MS: Absolutely. Look, like a lot of things in technology, we see a curve, right; and that cost curve comes down over time. It usually comes down as more and more people use the particular technology. So, you know, if you go back and think of the history of technology, we've seen this time and time again. What you really want to do, if you're in this particular business, is allow customers to use the technology. Establish those use cases and then, as you create scale, your costs come down. But there is an investment that's needed that's probably a forward investment, as Microsoft alluded on their recent conference call; and that shouldn't be terribly surprising. I had the opportunity last week, I was at a conference and I met with the chief operating officer of one of our holdings, and he really talked about the equation between the customer pain and what value this particular company was providing; and making sure that that equation was really a positive one for the customer, right? Because what you want to do is charge a price where companies really feel, or customers really feel, like they're getting a lot of value for it. And I thought he did a really good job of articulating the value. And he actually had a great use case with numbers and everything; and you could do the math, it was simple math. And I was just like, yeah, I would buy this, if I were running a business, because it was a positive net present value, i.e. it was a positive return. I don't want to try to talk in technical terms, but it was really interesting to hear him think about that. And I think that, you know, the great businesses - if you go back historically - they create value for their customers in using the product. So, you know, I mean, if don't you go by the wayside. So, I think that the business models will figure themselves out over time, but there probably is a required investment early on. As we think about artificial intelligence, and the applicability that we're seeing.

CR: From an investment point of view, presumably there are kind of two sides to it. There are those companies who are beneficiaries of AI and who will make money from it. And then there are the sort of disrupted companies who were kind of on the wrong side. And presumably you have to keep track of both of those. So, I wonder, can you talk me through companies that are benefiting and those that might be disrupted as well, that you're keeping an eye on?

MS: Yeah, I mean, in here again, you know, you have to make sure - we are living in a very quickly changing dynamic world, right? Where if you forego using certain technologies, if you're a business, you can get left by the wayside. Right, we've seen it with the movement of digital. We saw previously with the internet, etc, etc. So, if I think about artificial intelligence, it is another example. And I think the key point here is having data really creates an advantage. So, thinking about whether businesses that built their systems with data in mind from the get-go. I go back to earlier in my career - and I'll use Netflix for an example - I always thought that the beauty of Netflix was that they're just an amazing merchandiser of content, right? And it's because their system was designed from day one, with data - with the backend in mind, right? Like, what do they do with the data, that Mike Seidenberg who grabs a piece of content to watch? You know, how do they take that information and replicate it or use it across lots and lots of users in a very agnostic way, but to make sure that they're understanding not only how to get me to watch things, but also how should they purchase? What should they purchase? So here, again, I think in artificial intelligence, having that data and being data centric, is going to be really important to businesses. And if you're not, you're at a competitive disadvantage. So, we're spending lots of time as a team, thinking about those companies that basically have that data to leverage it into the artificial intelligence realm.

CR: Presumably, there is also a risk that share prices get ahead of the likely growth trajectory. I mean, there seems to be a sort of perennial problem with any fast-growing technology area. How do you manage that in the portfolio? And make sure you're not overpaying for that growth?

MS: Sure. You know, it is a perennial question as a technology investor. The notion of kind of thinking long term, yet also being really cognizant of the short term. And it really boils down to risk-reward. What is the risk-reward of a particular company, relative to their longer-term growth rate? Look, we manage it like a lot of people do. We very, very much think about position size, we think about when things seem too exuberant potentially, trimming that position. But it's active risk management, and it's a core tenet to how we run the portfolio. I will tell you with secular ideas - and I do think this is a secular idea - I'm probably going to err on the side of owning, versus not owning, given that in my career, I've just seen the secular themes play out really long and over many more years than investors typically realise. So, it's ideas, it's really identifying the companies and who we think the winners are, and going through our process and making sure we have exposure to those winners in some capacity.

CR: So just to wrap up, I mean, on balance, do you think this is an opportunity or a threat?

MS: I think on balance this is an opportunity. I mean, look, you know, I think that people thought - if you roll the clock back 70 years - there were huge pools of people typing in corporations, right? And those typing pools were replaced as the electric typewriter came of age, right? And everyone thought there'd be job loss, etc, etc. What you want, as the world progresses, is people using the best technologies to do their jobs in the most effective manner. Now, there may be some job loss, if your job is a very rudimentary one, where you just kind of go through the same task day in day out. But I think for a lot of people, I think that artificial intelligence is probably an enhancement, to the job opportunity for the average kind of knowledge worker. So, yeah, look, I don't think you can stop the progress of technology. There could be some job loss near term, but I think over time, it probably is more about job creation versus job loss. I will also caveat - it's early days - I think that there's a lot of learning to be done. But I don't think we're going to stop the wheels of progress. As people look at this technology, it’s such a key enabler to their businesses.

CR: Great. Okay, we'll wrap up there. If you have any questions on the subjects we've discussed today or anything about the Trust, please do go to the website - allianztechnologytrust.com or contact one of the sales team. So, thank you, Mike, for those insights today. And thank you all for listening.