Is it time to invest in (or run from) AI?

Still kicking yourself for missing an opportunity to invest in a company that would go on to change the world? It’s a rite of passage for many investors. Perhaps you wished you had owned Apple in the ’80s, Amazon in the ’90s or maybe Tesla in 2010, a few years after its first electric cars rolled off the assembly line. Investing can be hard, but missed opportunities can actually help make it easier for you to spot the next big thing.

With all the buzz around artificial intelligence (AI), you may be wondering if that next trailblazing company is already here. While some investors are eager to find a place for AI in their portfolios, privacy concerns about how personal data is accessed for this technology have somewhat tempered enthusiasm. Italy was the first western country to hit the pause button on ChatGPT by banning the processing of Italian users’ data. Italian regulators believe this technology may breach Europe’s strict privacy regulations. In North America, one of the early architects of Artificial Intelligence stepped away from his role at Google in May so he could speak more freely about his concerns about the safety of the technology. Despite the initial investor enthusiasm, this could be a good moment to pause and ask, “Should I invest in AI?”

If you’ve got questions about AI, well, so do we. The pressure to get in on any next-big-thing technology can feel very real. Here are some questions we’d want to cover before we consider investing.

What is AI, anyway?

Before you invest any money into this emerging space, it helps to understand what artificial intelligence is — and what it isn’t. AI is a broad term that gets used (and misused) in many different contexts. At a high level, AI is about developing computers to perform tasks that normally require human intelligence. This is what computer scientists refer to as “machine learning.” Much of this research pre-dates the personal computer, although it’s getting more attention today now that consumers can access software that incorporates AI.

“A lot of AI relates to teaching computers to recognize patterns, such as images of a cat, so they can learn and adapt from experience,” explains Jose Alancherry, Vice President, TD Asset Management. Once the computer learns the patterns, it can improve how it performs tasks and can even make predictions without additional programming.

So, what’s the difference between machine learning and deep learning?

The cat example is machine learning (or M.L.), a computer trained on a small set of structured data (images labelled as “cat”) that then creates its own algorithm to perform a task (predicting which images contain a cat). Machine learning requires human intervention to correct and learn.

Deep learning is the next evolution of machine learning. It uses neural networks — complex layers of algorithms and computing units that mimic the way a human brain’s neurons work — to process large amounts of unstructured data to learn on its own. It takes longer to train, but it can handle more complex problems.

The terms AI, machine learning and deep learning are often used interchangeably, but much of the exciting progress being made today is based on deep learning.

What makes chatbots like ChatGPT such a big deal?

ChatGPT is getting all the buzz because it’s the first tangible AI that consumers can tinker with, says Alancherry. Specifically, its ability to have human-like conversations has many people excited about its potential. According to an analysis by Swiss bank UBS, ChatGPT attracted 100 million active users within just two months of its launch in late 2022, making it the fastest-growing chatbot ever. ¹

ChatGPT means “chat generative pre-trained transformer.” In technical terms, it’s a large language model that can predict the next word in a series of words with realistic accuracy. It can create competently written texts of any form, including stories, songs, essays, emails, reports and programming code. Oh, and it can do it in almost any language. Vitali Mossounov, Global Technology Analyst at TD Asset Management recently spoke to MoneyTalk Live about recent developments in AI You can watch the interview here.

Sounds intriguing. What are some ways to invest in AI, then?

That’s sort of like asking, “How do I invest in the internet?” AI is not just one technology, and certainly not just one application. Having said that, one strategy could be to consider investments that feel the impact of AI expansion — both positively and negatively. By looking at existing companies that may play a secondary role in building AI technology, investors can identify possible opportunities. This is sometimes called a “picks and shovels” approach. (During the late-1800s Klondike gold rush, the merchants selling picks and shovels were often more successful than the gold prospectors.)

For example, some of the companies at the forefront of semiconductor chipset technology may benefit as more companies pursue AI because deep learning requires powerful graphic processing units (GPUs). “Any company that is involved in really advanced chip technology in the semiconductor supply chain could be one of the first beneficiaries,” says Alancherry. The same goes for companies that build supercomputer infrastructure, he notes.

Second-order benefits of AI may go to companies that harness its power to develop new software or hardware that dramatically improves work, health, or some other aspect of daily life. This is an aspect of AI that’s often exciting to think about, but more difficult to predict. One advantage investors have today over the tech leaps from 20 years ago is they can look at a smaller universe of companies to invest in. “Companies that have a clear technological edge today are fairly likely to maintain that at least in the near term,” says Alancherry. “You have a bit more of a quality buffer.” Still, he says, new companies could emerge that we haven’t even heard of yet.

How much of this is hype?

“No one company is the 800-pound giant in the room yet,” says Alancherry. “So, any investor should allocate risk capital accordingly.” There will also be companies that utilize AI in their operations to optimize processes and improve bottom-line profitability. Alancherry says this category may be harder to identify because they could choose to keep their competitive advantages a secret. Alancherry suggests investors benefit from keeping an open mind as to AI’s potentially transformative effects on the way businesses are run.

In broad terms, there will be companies where AI is the product and others where AI will power the product. In either situation, Alancherry urges investors to take their time to really understand how companies are using the technology. The innovative companies will quickly scale up, but it’s important to look for tangible proof, and avoid relying on others’ opinions alone.

What industries will feel the impact of AI first?

AI could eventually affect many or even most industries. “These tools are really good at repetitive processes that are often mundane to a lot of humans,” says Alancherry. He points to health care as one field with some of the fastest, most tangible applications for AI.

“Finding new drugs is a matter of trial and error, and a lot of the time, it’s just running extensive tests and studies to evaluate combinations of molecules,” he explains. AI can significantly accelerate this process by simulating outcomes and narrowing the focus of work. “AI allows them to eliminate a lot of the dead ends and get products to market faster. Companies at the forefront of medical innovation stand to directly benefit from AI” For instance, some pharmaceutical companies are looking at AI in drug discovery and biologics, while others are being used to watch for subtle changes in a patient’s vitals to alert doctors to a potential health condition.

What can previous tech booms teach us?

Beware of the hype cycle, says Alancherry. “It’s common for there to be an over-expectation that a new technology will transform the world overnight,” he says. Often, investors are let down when it doesn’t happen. “But we also underestimate the change in the long term. So, in the background, the technology continues to progress.”

How a company ultimately turns technology into a viable business is about much more than the technology itself. The transition phase is always messy, so there will be displacement and disruption. It’s hard to completely visualize all the opportunities to come.

There is plenty of time to still explore avenues to invest in AI “We’re in that initial phase where everybody’s excited, but the process will still be gradual,” says Alancherry. “These are technological frontiers, and it will take time for them to fully crystallize into everyday use.”

What does ChatGPT tell us?

Since ChatGPT was created to answer questions, we wondered what would happen if we asked it, “How can I invest in artificial intelligence?” Here is its response:

     Investing in AI can be lucrative opportunity for those interested in emerging technologies.
     Conduct thorough research on companies involved in AI development.
     Look for companies with a strong track record and promising future prospects.
     Consider potential risks and market volatility associated with this emerging technology.

That’s good advice for any investment, and when it comes to AI, that’s only part of the story. AI is only as good as the information it can find online. If you want to enter this space ahead of the crowd, you’ll still need to do your own human-powered research.