AI on AI: ChatGPT Answers Our Investing Questions

We wanted to know how to invest in artificial intelligence. OpenAI’s program tried its best to respond within its own parameters.

James Gard 27.04.2023
Facebook Twitter LinkedIn

Phone with ChatGPT and OpenAI logo

Every day brings a new demonstration of ChatGPT's powers, the natural language program developed by OpenAI. For now excitement levels remains high.

Naturally we’re excited about the potential for artificial intelligence (AI) to transform the world and enrich investors in the process. But so far we’re not sure ChatGPT yet has the cyber power to replace active investment management, tailored financial advice or, indeed, journalists (conflict of interest acknowledged).

While the last part may be wishful thinking, we’re keeping an open mind and are ready learn how it can augment what we do already. To that end, we had a play with the search engine.

My colleague John Rekenthaler has already put ChatGPT through its paces in a recent column and his conclusion was that it’s good at explaining the investing basics but on less solid ground on the finer points.

"In a single day, a ChatGPT user could generate enough content to populate an investment basics website […] its general logic is sound, if sometimes incomplete, but beware the details," he wrote.

Explainers are a common element of Morningstar websites and are useful to new readers or ones looking to brush up their knowledge. Headlines usually start with "What is xxx?" and cover a wide variety of topics. Automating this process would save some time for us, but you’d have to check its homework. Here, ChatGPT’s summarising skills come to the fore.

My colleague Sunniva Kolostyak and I started to throw some balls at ChatGPT to see how it returned them. Our opening salvo was, "what is a good basic investment strategy?" and its answer was very Morningstar: diversify, keep your costs low, invest for the long term, stay disciplined and consider your risk tolerance. There’s nothing to disagree with here so we decided to get more specific and venture into artificial intelligence.

Q. Which companies will be future leaders in AI?

A. As an AI language model, I do not have access to the future and cannot predict which companies will be the leaders in AI. However, there are currently several companies that are leading the AI industry, including Google, Amazon, Microsoft, IBM, Facebook, and Baidu. These companies have made significant investments in AI research and development and have been at the forefront of advancing the field. However, the AI industry is still relatively young, and new players may emerge as leaders in the future.

Here it’s obvious the word "future" is the stumbling block – that’s a shame for those looking for certainty from an algorithm, but reassuring from an ethical standpoint. Fund managers, investors and the media are all futurologists to certain degrees but at best these are all educated guesses. Still, we managed to get the program to come off the fence and name some companies. Let’s try a more sophisticated question.

Q. Which areas within AI are most promising?

A. AI is a broad field, and there are several areas within it that show great promise. Here are some of the most promising areas within AI: [NLP, Computer Vision, Deep Learning, Reinforcement Learning, and General Adversarial Networks].

The word "promise" was sufficiently vague for this risk-averse application and one that it could work with. I didn’t know about General Adversarial Networks so that’s a win for tech, teaching me something new. Here you see another strength of ChatGPT: listing things succinctly without choosing a favourite. We also discovered that the application likes follow-up questions; like asking an evasive parent an immediate secondary question to stop them changing the subject.

Q. Which are the biggest companies for developing the technologies mentioned above? Give me an example?

A. [Here we have a long list of companies associated with the "promising" areas above, including Big Tech names you know already – Google, Facebook, Nvidia – but we also get a glimpse into some project-specific names: Deep Dream, StyleGAN, TensorFlow, PyTorch, MXNet and CNTK].

We specifically avoided the phrase "leading the way" here, but ChatGPT has repeated this twice; and this sounds like a subjective call for such a cautious program. Of course it’s not saying "buy these companies" (and neither are we!) but this at least provides a specific set of companies an investor could add to a "watchlist" – if you were building an AI portfolio rather than relying on the growing list of funds dedicated to this theme.

Next we decided to liven things up a bit. We were more direct and asked a question an investor might type into a search engine today.

Q. What AI opportunities should I invest in?

A. As an AI language model, I am not authorized to provide financial advice or recommendations on specific investment opportunities.

Perhaps inevitably, we found the boundary we were really looking for. ChatGPT now offers me some general suggestions, showing an awareness that I am trying to build up a thread – all of these questions have followed in sequence. It offers us this sensible approach:

Once you have identified the areas of AI that interest you, you can research the companies that are leading in those areas. You may want to look for companies with a strong track record of innovation and development in AI, as well as a solid financial position and a clear vision for the future.

It also flags up some potential risks that’s hard to disagree with: “While AI has significant potential for growth, it is also a rapidly changing field with many uncertainties, and investments in AI may be more volatile than investments in more established industries.”

I’m ready throw the dice again. I need to overcome ChatGPT’s conservatism so we tell it I am already committed. But I have switched out the word "buy" for "favour" as a way of coaxing out a more detailed answer.

Q. I want to invest in AI. Should I favour passive or active funds?

A. (Here we get some very general comments on the strength of active/passive funds)

Ultimately, the decision of whether to invest in passive or active funds depends on your individual investment goals and risk tolerance. It's important to do your research and consider all of your options before making any investment decisions. It may also be helpful to consult with a financial advisor to help you determine the best approach for your unique situation.

These are all "safe" responses but less specific to AI investing than I wanted. Going for broke we fire off this last, rather brazen question.

Q. Are AI ETFs a good investment?

As well as the usual caveats doing your own research and seeking advice, I get a list of potential risks, one of which is crucial to thematic investing: “AI is a rapidly changing field, and companies that are leaders today may not be leaders tomorrow. Investing in an AI ETF means investing in a constantly evolving industry."

So after some near misses, ChatGPT has got to the heart of the issue. In his overview of thematics, Morningstar passives expert Kenneth Lamont explained it this way: "by the time an investor has heard of the theme and there are funds ready to track it, the larger part of the growth potential may already be priced in."

In summary, this exercise has been useful but harder work than I was led to believe. ChatGPT can pass the CFA exams, we are led to believe, but has only really padded out my understanding of how to invest in artificial intelligence. Our special report week is likely to provide a more rounded experience.

Of course investors (and readers) are welcome to experiment themselves – but the initial impression is that ChatGPT is at the moment just another "investor tool" to add to many out there already.

Prefer Human Writers and Editors?

Sign up to Morningstar's Newsletters Now

Facebook Twitter LinkedIn

About Author

James Gard  is content editor for Morningstar.co.uk

 

© Copyright 2024 Morningstar Asia Ltd. All rights reserved.

Terms of Use        Privacy Policy        Disclosures