Page 23 - Investment Advisor July/August 2023

gies, which include machine learning, robotics, automation, fit the theme,” Armour said.
speech recognition, natural language processing, computer The exposure may be indirect, but so far this year, AI ETFs
vision and artificial neural networks, among others. have posted strong returns “largely because they own the
Six of Morningstar’s seven identified AI ETFs are passively kinds of stocks that people would not be surprised to see doing
managed index funds. VettaFi has identified a longer list of well,” including NVIDIA, Intuitive Surgical, Amazon and
AI ETFs. Microsoft, Nadig said.
Investors may seek purer exposure to AI with a portfolio of
Investment risks, Challenges more AI-centric companies rather than trillion-dollar multi-
AI ETFs have relatively short track records and almost all nationals with many other revenue streams, Armour said, not-
have small net assets, except for BOTZ, which is heavily tilted ing that investors often already have exposure to the FAANG
to robotics, Armour noted. Investors should look at the hold- stocks (Facebook parent Meta, Amazon, Apple, Netflix and
ings to see if they align with why they’re interested in AI, and Google parent Alphabet) through index funds.
consider fund fees, he said. “I would look for companies that I see as potential AI lead-
It’s also critical to understand how a more volatile tech ers or companies that could really benefit from AI more so
exposure fits within portfolios, Armour added. “You do tend than the broader market and then I would see which ETF sort
to take on a pretty concentrated risk exposure which can defi- of fits that roster best,” Armour said.
nitely be more risky,” he said. VettaFi’s Nadig noted that pure-play investing is often
The risk level is moderate, however, if the AI ETF repre- more volatile, with smaller-cap companies bringing higher
sents a small place in a portfolio, Armour added. “This should risk. “Those are the tradeoffs that you’re going to have to
not be a part of your core holdings,” even for investors with make,” he said.
a high risk tolerance, Armour said. “I
don’t think you need to go all in on one
of these ETFs.” Key AI Terms to Know
“You have to identify the right theme Conversational AI: Technologies that use large volumes
and then the ETFs have to provide of data, machine learning and natural language process-
exposure to that theme in the right way ing to allow users to “talk to” the technology, by imitating
and then you have to be buying in at the human interaction through recognizing text and speech inputs.
right price,” Morningstar’s Armour said. Conversational AI serves as the synthetic “brain” behind some chatbots.
“AI seems like a durable theme right
now, so I think it’s fair to be interested Generative AI: A category of AI systems, including large language models, that
in investing in AI and trying to figure can independently create unique, novel content in the form of text, images, audio
out how to do that.” and more, based on the data they have previously been trained on.
The greater challenge is figuring how
artificial intelligence ETFs are providing GPT: generative pre-trained Transformer; the prefix to various generations of
exposure to AI, Armour added. “AI is large language models from the company OpenAI. For example, gpT-3 is the third
fast-moving and it’s hard to pick which generation of gpT models. gpT-1 was released in June 2018. gpT-2 was released
companies properly reflect AI or how to in February 2019. gpT-3 was released in June 2020. gpT-3.5 was released in
look at AI,” he said. March 2022, with underlying models rolled out over the year. gpT-4 was released
While AI funds invest heavily in some on March 14, 2023.
of the world’s biggest firms — NVIDIA,
Amazon, Microsoft (which has invested Machine Learning: A broad branch of AI concerned with “teaching” AI systems
billions in ChatGPT developer OpenAI) to perform tasks, understand concepts or solve problems in a way that imitates
and Apple — “you wonder how much of intelligent human behavior, gradually becoming more accurate as it is trained on
the revenues coming out of these hold- more data.
ings are going to be directly tied to AI,”
Armour said, noting that the big players robotic Process Automation (rPA): A form of business process automation, also
are involved in many other products known as software robotics, that allows humans to define a set of instructions
and services. for the performance of high-volume, repetitive human tasks quickly and without
“That’s not going to be direct expo- error. While RpA technology shares similarities with AI, it is not a form of AI.
sure to AI that you’re getting. So it’s —Beth Braverman, rhys Dipshan and stephanie Wilkins
tough to pick out which ETF might best

July/August 2023 InvesTmenT ADvIsor 21