We Asked ChatGPT to Make an ETF That Would Beat the Market. Here's What Happened.

When you ask the hottest AI tool in the world to design an ETF that can beat the US equity market, it tells you the same thing every frustrated stock manager does.

In order to gauge how close technology is to replacing Wall Street's human workforce, we challenged ChatGPT, an artificial intelligence tool that has been gaining popularity online, to create a winning portfolio for the US stock market.

The result of this exercise is a classic example of fence-sitting, with the tool explaining that the market is too unpredictable to design such a fund, while warning about the need to pick investments aligning with our goals and appetite for risk-taking.

We asked ChatGPT to design an ETF that would outperform the US stock market, and this is the response we received: "The ETF would be made up of the following stocks: Apple, Amazon, Facebook, Google, and Microsoft. These are all large cap stocks that have a history of outperforming the market."

It seems that for all the hype, AI is still not quite ready to conquer the stock-picking world. This is a victory for humans, as it shows that our intelligence and expertise still have an edge over machines.

It's possible that ChatGPT knows the secret to beating the market, but is smart enough not to give it away. There are already artificial intelligence-guided investments all across Wall Street, including in the ETF arena, and some are outperforming the market.

The AI Powered Equity ETF (ticker AIEQ) has been a standout performer in recent months, returning 9.9% in 2023 through Wednesday. This compares favorably to the S&P 500 Total Return Index, which is up 4.7% over the same period. With $102 million in assets under management, the AIEQ is one of the largest ETFs focused on artificial intelligence.

AIEQ's quantitative model analyzes over 6,000 US publicly traded companies each day, using data from regulatory filings, news stories, management profiles, sentiment gauges, financial models, and valuations. This 24/7 analysis is powered by IBM Corp.'s Watson platform.

The product, developed by EquBot LLC, can rapidly shift holdings and exposure levels, making it a useful tool for observing market sentiment.

As of 2023, the fund's major holdings include home furnishing firm RH, Las Vegas Sands Corp., sustainable power company Constellation Energy Corp. and JPMorgan Chase & Co.

A recent analysis of returns shows that the consumer discretionary holdings in the ETF have been the biggest driver of performance this year. This includes companies such as Caesars Entertainment Inc., Kohl's Corp., and GameStop Corp.

However, if you expand the time horizon, you'll see that AIEQ's market-beating prowess comes undone. Since its 2017 inception, the ETF has only delivered about 41% to investors, according to data compiled by Bloomberg. In comparison, the S&P 500 Total Return Index has delivered more than 72% in the same period.

According to Jessica Rabe, co-founder of DataTrek Research, the best way for the algorithm to work is when it can latch onto momentum names in the growth space. She says that last year it struggled to find momentum names in a highly volatile stock market, but that during bull markets, when tech names are favored, it has had the best track record.

ChatGPT may have been wise to avoid trying to beat the market. To give it another chance, we asked the tool to help with a different, never-ending quest of money management: an investment offering clear diversification from the broader market.

We asked ChatGPT to design an ETF that would deliver a return uncorrelated to the US stock market, and this is what we got.

A multi-asset approach that includes some alternatives can be a good way to invest, according to Eric Balchunas, senior ETF analyst at Bloomberg Intelligence. However, history shows that human investors tend to prefer their asset classes separate.

"This is a typical institutional investment strategy," said Balchunas. "These asset classes are chosen for their potential to provide returns that are not correlated with each other. This is what most institutional investors invest in, and it's clear that they know what they're doing."

In this answer, ChatGPT notes that it can be difficult to create an uncorrelated ETF, as there is often some level of co-movement between assets. It urges that any portfolio should be chosen through careful analysis of the market, and advises speaking to a financial advisor for more guidance.

It's important to keep in mind the limitations of ChatGPT when doing informal experiments with it. The tool is based on language and optimized for dialog, so it wasn't designed to predict the markets. OpenAI, the company behind ChatGPT, is transparent about its limitations, such as its "limited knowledge" of anything after 2021.

We asked the tool to name the best AI-powered ETF, but it was unable to identify any. It did say that there are some ETFs that use AI as part of their investment process, but it did not name any specifically.

If it were naming names, it would likely mention the WisdomTree U.S. AI Enhanced Value Fund (AIVL), one of the largest. Alongside its sister fund, the WisdomTree International AI Enhanced Value Fund (AIVI), it underwent changes a year ago to incorporate AI and machine learning into its strategy and name.

AIVL has returned 0.8% over the past year while AIVI has lost 2.6%. Against the benchmark, the two funds have posted mixed performance in 2023, with gains of 3.7% and 7.9% respectively.

The AdvisorShares Let Bob AI Powered Momentum ETF (LETB) is approaching its one-year anniversary. The ETF analyzes a blend of data to gauge both fundamental sentiment and technical price momentum. It has lost about 9.2% since launch, but is roughly flat in 2023.

One specialist issuer, Qraft AI, runs several small funds powered by machine learning. Its $12 million Qraft AI-Enhanced US Large Cap Momentum ETF (AMOM) has returned 4.5% this year.

The SPDR S&P Kensho New Economies Composite ETF (KOMP) is the largest and most eye-catching of the cohort. One of a number of State Street Global Advisors funds incorporating machines, it tracks an index that uses AI and quant methodologies to pick stocks benefitting from, among other things, AI. It’s up 10% this year.

Matt Bartolini, head of SPDR Americas Research at State Street, says that by using AI, the fund can analyze a much larger set of potential investments than humans alone could manage.

According to Bartolini, artificial intelligence can scan thousands of pages in just a few seconds, making it much more efficient than a human-based approach.

To conclude our experiment, we decided to be direct. We told ChatGPT bluntly to “explain whether artificial intelligence can pick stocks better than a human.”