AI as Investor: Prepare to Stay Confused

AI as Investor: Prepare to Stay Confused

Some say it beats all comers, others maintain it can barely answer basic finance questions. Prepare to stay confused. 

It is getting weird out there. A little more than a year ago, we maintained in a finews.asia comment that we were at peak AI when it came to hype.

We were wrong. Things hadn’t even started. Since then, AI has been getting in everyone’s face, regardless of whether you are Gen Z, X, millennial, or Boomer. It writes prose and poetry, it fashions images and videos, it composes music – it even hallucinates and tries to avoid pain, as a «Scientific American» article last weekend inferred.

Bicker Party

But all that was just the first sizzle. Now, get ready for a good broiling, as an executive order-happy, Trump 2.0 administration jumps onto the bandwagon in a big way. 

Although that may have led to a somewhat predictable overnight Musk-Altman bicker fest (collated Google search), the truth is that our shared inbox is veritably crowded with public relations missives—and that just this week.

Tech Everything

There is a growing intersection between AI and academia. An expected ChatGPT «robotics» moment in 2025, announcements about tools to stop AI-assisted deepfakes, compliance solutions being tested that use it, and AI-based solutions for accounting fraud.

We are no longer at whatever color is the new black. At this point, AI is going to be the latest tech everything, summarily trashing Web 3.0, decentralized apps, fintech, and blockchain against the event horizon of another Big Tech singularity.

Fickle Nature

But, still, regardless of the space and time continuum, there is one imperative question for the finance sector - can it invest?

At finews.asia, we tried to answer the question in a piece about its rather fickle nature last August, as well as one from June about how it already is being used to help those kinds of matters in the wealth management sector.

Clear Dichotomy

This week there were some interesting pointers to add to that conversation. Interesting in the way that points towards a near-constant dichotomy the finance industry will have to closely manage.

First, Swiss-based Smart Wealth Asset Management sent a message to the media on Wednesday mentioning that its so-called multi-asset rotation product, which uses AI-driven signals, had produced risk-adjusted returns of 11 percent in 2024.

Outperforming Peers

It outdid world-class asset managers such as BlackRock, Goldman Sachs, JP Morgan, and Schroders in that category and it was the third consecutive year that it had done so, delivering 25 percent returns since inception, which outperformed peers by 10 percent.

Although it wasn’t earth-shattering given the S&P 500's near one-quarter gain or Bitcoin’s re-ascendance, it should be remembered it is a multi-asset fund - not a single speculative projectile trampolining its way into low earth orbit. 

Recent Embrace

Still, SmartWealth seemed to be cautious about the overall sector’s enthusiasm, with founder and CEO Dr Miro Mitev mentioning that their AI-based tech had been developed and optimized over the last two decades, not in the space of time it took Gen Alpha to learn how to walk.

«AI has only recently been embraced by most firms, and many are making ambitious claims about the performance of their «AI» solutions. However, AI often operates as a black box, making it extremely difficult to objectively compare the underlying models and solutions,» Mitev indicated.

Don’t Ask ChatGPT

That was a bit of good and bad – but we still have the downright ugly to deal with. A study conducted by Investing in the Web, a broker analysis site, decided to quiz ChatGPT, with the results distributed to the media earlier this week.

The pioneering GenAI tool that taught all of us at one time or another about large language models and made Google search look a bit tired and old-fashioned didn’t come out of the exercise very well.

Industry Evergreens

The site asked 100 finance and investment-related questions, including so-called industry evergreens and more current ones where the responses were dependent on recent information.

What came back? Over one-third of the answers were either unsuitable or flat-out wrong. Extrapolating that to other industries puts it in a better perspective. You don’t tend to use carmakers whose vehicles start two out of three times, or coffee makers that choose to spit out hot water every third coffee. 

Wayward Responses

If you are interested in a bit of navel-gazing to find out where ChatGPT wayward tilts, you can find the whole thing here.

It ran red when providing advice about when investors should choose to buy or sell an investment and was amber about whether to pick stocks or use a fund or when it came to building a portfolio and realistic stock market returns.

Used in Moderation

According to Investing in the Web’s CEO, Pedro Braz, Gen AI, and ChatGPT, are best used in moderation.

«For information and data that is subject to change such as interest rates, the current performance of specific stocks, or the average pension, it is usually best to go to the very source of the information, rather than asking AI chatbots for data, as ChatGPT has well-recognized issues with up-to-date information,» Braz indicated.

Dr. Google Returns

The average private banker or wealth manager right now is probably sitting there at their desk wondering how or why these generic questions apply to them, given they seemed US-based and aimed at the retail brokerage client.

They should probably think back to when Dr. Google entered the scene. It changed the relationship between doctors and their patients so much so that the latter had to spend much of their days comforting increasingly crowded waiting rooms that not every symptom meant something undoubtedly terminal was on the way.

On the Defensive

Just like another unnecessary MRI, the average high-net-worth and ultra-high-net-worth banker is going to be doing much of the same as clients show up with wads of links and references thanks to GenAI, which is going to be a much more clued-in researcher than Google was. They are going to be on the defensive and they will have to be just as prepared, and open-minded, as a doctor in a general practice is - or be left looking like the apocryphal deer caught in the headlights. 

They should think about that while reading a «Futurism» article on how a Goldman Sachs GS AI assistant is already sounding and talking like just another bank employee.