Diversified access to artificial intelligence: Invesco Morningstar Global Next Gen AI Index ETF

00:00:04:29 - 00:00:31:08

Speaker 1

Hello, my name is Marcus Berry, and I'm an ETF specialist at Invesco. In this session, we'll be discussing artificial intelligence, or AI, with the aim of providing some thought leadership around the subject and also considering the investment case for AI. Over the past 18 months, artificial intelligence has arguably become one of the most talked about and important trends, not just within the investment industry but also within the mainstream media.

00:00:31:11 - 00:00:54:25

Speaker 1

There's a lot of excitement about the potential of AI and the impact it may have on the economy and our lives. But there are also some valid concerns too. In order to help you better understand the AI opportunity and have better conversations with your clients, we’ll look to answer three questions. Firstly, what is artificial intelligence, and why has it been such a sudden excitement within the topic?

00:00:54:28 - 00:01:24:04

Speaker 1

Secondly, how might AI impact specific sectors of the economy in the market? And lastly, how might investors consider getting access to the AI theme? To help me answer these questions, I'm pleased to be joined by my colleague Ashley Oerth, who works on the Invesco Global Thought Leadership team in her role as a senior investment strategy analyst. Ashley recently wrote a series of white papers on AI, and those are available from your Invesco representative.

00:01:24:06 - 00:01:31:09

Speaker 1

So, Ashley, let's kick things off at the very top. Can you provide an overview of AI and what it is?

00:01:31:12 - 00:01:55:30

Speaker 2

Sure. And very happy to be here, Marcus. Thanks so much for having me on. So, about AI. So, as I'm sure everyone is already aware, AI really took center stage in the media last year with the advent of ChatGPT, and this sort of came seemingly out of nowhere for many of us. So, I think it's a really good idea to take a moment to really clarify it.

00:01:56:00 - 00:02:16:04

Speaker 2

And AI, in its most basic form, it's really about mimicking some kind of human intelligence or decision-making. It can, for example, help us process and categorize data, make decisions based on available data, or even create new data based on a prompt, which is what we've seen more recently.

00:02:16:04 - 00:02:24:22

Speaker 2

And so, while the world of science fiction really gives us fantastic and scary ideas of what AI may be today, it's really just a tool.

00:02:24:28 - 00:02:47:18

Speaker 2

In fact, AI is a term. It's generally used to refer to as systems that quote-unquote learn from data either with some kind of target purpose in mind or understanding linkages within data. And so, while it may feel quite new in some ways, in reality, it's something that's been around for quite some time and is already broadly deployed in today's economy.

00:02:47:20 - 00:02:56:20

Speaker 1

Okay. So, you mentioned that it's been around a while. Why the sudden excitement in the last 18 months or so?

00:02:56:22 - 00:02:57:20

Speaker 2

Sure.

00:02:57:20 - 00:03:22:23

Speaker 2

And I think that this sudden excitement, as you've described it, it's really because of the advent of generative AI. I mentioned ChatGPT just a moment ago. This is one example of it. And, you know, as I already mentioned, AI has been around for quite some time. But it wasn't until this past decade that we began to see meaningful results from this sort of subset of AI that we refer to as generative AI.

00:03:22:26 - 00:03:50:12

Speaker 2

And I would argue that three factors have sort of helped enable this in this present moment to arrive today. So those three factors are, number one, that we have more data available than ever before -- that more data is generated today in a single hour than an entire year two decades ago. And data, it's used to enrich models, and it's really the essential ingredient to improving their outcomes.

00:03:50:15 - 00:04:20:16

Speaker 2

Then the second factor is that we have greater computing power available to us. The computing power over the course of many years has continued to grow, including more sophisticated hardware and software as well for distributed computing tasks. And this greater scale that this allows really shows up in cloud computing operations. And all of this, it's essential to allowing AI systems to really harness ever more data by providing the computing resources for it to really work through these.

00:04:20:19 - 00:04:44:17

Speaker 2

And you can think of them as piles of data. And then the third factor is that over the same period, we've also seen incremental model improvements. The models have improved so much with new methods that are really helping to pioneer more intelligent systems that can appreciate context, that together with, as I mentioned, more data and more computing power, really enable this sort of AI moment today.

00:04:44:19 - 00:05:22:13

Speaker 2

And so, with all these advancements, now we see these generative AI systems able to, you know, respond and in text or generate images for us. And they do this with capabilities that are now like human levels or human capacities. So ChatGPT -- really one of the most highly publicized gen AI tools to be released recently -- it's been able, you know, with all this background, to pass things like difficult law exams and graduate school entrance exams and many other examples where really complicated tasks, it's able to solve for and overcome.

00:05:22:16 - 00:05:32:03

Speaker 2

And so, these capabilities, they're impressive. And now commentary has really centered on how generative AI may transform our working lives.

00:05:32:05 - 00:05:48:15

Speaker 1

Fantastic. Thanks for providing that background around generative AI and why we've seen such excitement over the past 18 months. But could you also provide some potential use cases or applications for how generative AI can be used in, you know, in our lives today?

00:05:48:18 - 00:05:55:06

Speaker 2

Absolutely. So, as I mentioned, AI, it already exists in so many forms already in today's economy.

00:05:55:06 - 00:06:26:22

Speaker 2

And, you know, there's a variety of technologies that we take for granted on a day-to-day basis, everything from voice assistants to content recommendation, spam filters, facial recognition. Even just using a maps application. And all of these are just examples of the AI that came before the generative stuff that we see more recently. Now, with generative AI, it’s sort of like we've been given kind of like a lump of Play-Doh or a box of Legos that we can use to create really very many things.

00:06:26:27 - 00:06:50:26

Speaker 2

And I think at this stage, it's really about taking these capabilities that we see today and being able to specialize them, that we're applying these quote-unquote foundation models, these sort of ready-made packages that are ready to be specialized for a more refined use case. And it's in this process of specialization that we can find really more interesting use cases.

00:06:50:29 - 00:07:30:08

Speaker 2

One of my favorite examples really comes from the field of medicine, where generative AI can rapidly test new ideas for drugs based on the sort of biochemistry grammar that is able to be learned much in a similar way as we can teach an AI system to process and generate language. So, in the same way, we can, you know, chat with a chatbot that is powered by generative AI, we can do something very similar, except that instead of producing text, we're instead producing novel proteins that can be tested in medical contexts and help enable research breakthroughs.

00:07:30:14 - 00:08:02:05

Speaker 2

So, really, I think this is one example of a list of many to show that we can use this technology for accelerating research and development. And in fact, I really think of generative AI as being sort of a good candidate for anything that involves lots of data. You know, for example, there's many models on market today that I can pull up, give it a PDF document that I want to read, and it can summarize it, and I can interact with it to ask it, you know, on what page is this topic discussed so I can really dig into the details where I need to.

00:08:06:05 - 00:08:23:20

Speaker 2

And so, there's these examples that can really help us, you know, be more productive in our jobs, in research and again, in more information, heavy tasks. And I expect that as the business world becomes more familiar with AI and its capabilities, that there's likely to be ever more such applications.

00:08:25:16 - 00:08:47:21

Speaker 1

Okay. Well, there's a lot of excitement about AI and generative AI today, not just from the general public but also within the stock markets itself. A lot of investors are excited about the potential applications. Could you provide your thoughts maybe around some of the wider benefits and opportunities for the economy?

00:08:47:24 - 00:08:49:18

Speaker 2

Absolutely.

00:08:49:18 - 00:09:14:15

Speaker 2

And, you know, I think that I already touched on some examples. And really, what this gets at is that generative AI, if we're able to harness it in our day-to-day work lives, it may very well increase productivity growth and drive higher real incomes and maybe even help exert a deflationary force on the economy. We already know that technology has a pretty long history of, you know, having a kind of deflationary effect.

00:09:14:17 - 00:09:47:29

Speaker 2

AI may be another driver of that. So, in essence, I sort of view AI or generative AI, more specifically, as being a tool that can help us free up valuable work hours for other higher value add tasks. My favorite example comes from software engineering, where there is a study that was conducted that paired a number of software developers with an AI coding assistant, and there is another, you know, a test group or a control group, excuse me, that had no such assistant.

00:09:48:02 - 00:10:22:10

Speaker 2

And in this example, they actually found that those software engineers that were paired with an AI assistant were able to complete this task 56% faster than the control group. And so this really gets at the idea that AI can help us be more productive in our day-to-day jobs. And if it's the case that we can see this in a broad experience across the economy, this may well help it reignite productivity growth, which, you know, economy wide and across the world even has been on a quite dismal trajectory.

00:10:22:11 - 00:10:46:23

Speaker 2

So, you know, if we're able to see a revival similar to the period 1995 to 2005, this is the period of the information and communication technology revolution, or the ICT revolution, that we would be able to see productivity growth maybe three times higher than what it's been in the past decade. And this, in turn, should help drive, again, higher real incomes, lower prices, and ultimately larger economies.

00:10:46:23 - 00:11:02:09

Speaker 1

One of the areas where I think AI is divided opinion is on its eventual impact on the labor market. You know, who is going to be the winners and the losers? And I know this is something you touched on in your white paper series. So, can you provide your thoughts on that?

00:11:02:20 - 00:11:26:03

Speaker 2

Sure. I think that this is a natural concern for people to be having that, ultimately, people fear being replaced, and if AI does have these sorts of human-level capabilities, that therefore that may put in jeopardy some jobs. And I think that's, you know, that this is a valid concern, and I have a lot of sympathy for it. But with that said, I'm not too worried.

00:11:26:10 - 00:12:07:29

Speaker 2

So, despite, again, these years of advancing advancements in AI that we've already seen, which includes widespread deployments across commerce, applications, and more, we have one of the tightest labor markets still on record. And on top of that, AI, it's still prone to making mistakes. That means it needs close human supervision. So, AI, it's something that I would argue has more specialized use cases, that it's not something that can necessarily, you know, obviate the need for a human in a role or in the loop of a task or process.

00:12:08:01 - 00:12:41:08

Speaker 2

Rather, it's more adept at handling particular tasks, especially information-heavy ones, sort of like, you know, finding the needle in the haystack, if you will. But still, maybe, you know, to continue that sort of metaphor that you would have somebody who would first identify what that needle looks like. So, I think that's the human role remains. And with all that said, AI is still something that's expensive to train, expensive to run and maintain, and it requires quite expensive talent to really produce and keep models running as expected.

00:12:41:11 - 00:13:03:17

Speaker 2

So, this means that there is running costs that demand investment in our entire supplies for semiconductors. And that, with all that said, would suggest to me that AI will be applied to the places in which it is most cost-efficient to do so. So I expect I will really be a harness where it is most cost-effective, you know, across information-heavy tests especially.

00:13:03:23 - 00:13:16:21

Speaker 2

So, this would leave, in my view, a substantial role for human labor. And so, again, I'm not too concerned about this sort of story of AI automation that's impacting the labor market.

00:13:17:04 - 00:13:30:24

Speaker 1

Fantastic. Ashely, thank you so much for your comments there and answering those questions. And I just want to maybe answer that third question that I posed at the introduction, which is: How can investors gain access to this theme?

00:13:31:02 - 00:13:57:21

Speaker 1

Now, clearly, while some investors may want to buy individual securities, we believe as an ETF provider that one of the more efficient ways to gain access to a theme like AI is to buy a diversified portfolio. And we're very pleased and excited that Invesco, in January 2024, we launched the Invesco Morningstar NexGen AI ETF, ticker INAI.

00:13:57:23 - 00:14:08:24

Speaker 1

This is an ETF that provides exposure to 50 names across the AI theme and provides exposure to four main areas within AI.

00:14:08:24 - 00:14:19:08

Speaker 1

First of all, generative AI. Then, you have data and infrastructure, AI software, and AI services.

00:14:19:11 - 00:14:46:00

Speaker 1

And if you look at the top 20 names in the portfolio, there may be some familiar names there, like NVIDIA, Microsoft, and Alphabet. But beyond that, you get exposure to some of the more up-and-coming AI names that may be not as common in investors’ portfolios, like a Palantir Technologies, a Snowflake, or a Baidu. If you're interested in getting more information about this ETF, please reach out to your investor representatives.

00:14:46:03 - 00:14:52:19

Speaker 1

Again, the ticker symbol is INAI and with a management fee of 0.35%.

00:14:52:19 - 00:15:03:23

Speaker 1

Thanks again for taking the time to listen to this video. If you want any more information, again, please reach out to your Invesco representative or visit the Invesco website. Thank you.