I Let AI and My 5-Year-Old Pick My Stocks. Who Did Better?

I consider myself a disciplined investor. I stick to dollar-cost averaging and dividend reinvestment plans. I sift through quarterly earnings reports; I generally avoid meme stocks and IPOs no matter how much hype they receive. And I pride myself on having a well-diversified portfolio.

Is it boring? Sure. But is it effective? Absolutely.

It was for my grandfather, too, who taught me the basics of investing, including how patience is as critical to long-term success as which companies I choose to purchase shares of. He passed away in 2006, and 18 years later, my grandmother — who turned 92 recently — is still living comfortably from the seeds he sowed decades ago.

But when a bunch of self-proclaimed apes begin to target short sellers, or when a new technology claims it can produce returns that beat the market — a feat that the majority of investment professionals are incapable of achieving — it always piques my interest. Such was the case earlier this year when a company named Danelfin began claiming that its AI-powered stock-picking tool was capable of outperforming the benchmark S&P 500 index over 90 consecutive trading days (i.e., excluding stock market holidays and weekends).

It was a bold assertion made by the little-known startup, and as Money's investing editor, I would've been remiss not to have tried it out for myself. So in short order, I signed up for Danelfin's free model and decided to pit it against the S&P 500. And while I was at it, I asked my son — whose interests currently revolve around Roblox and ice cream — to hit the big blue button on a digital stock randomizer until he got bored.

The stage was set: the S&P 500 vs. AI vs. my 5-year-old.

How Danelfin's AI works

Before diving into the stocks chosen for me, it's useful to understand how some companies are leveraging AI to help improve investors' odds of success.

The ability to beat the market isn't taken lightly in investing circles. BNP Paribas found that in 2023, hedge funds returned an average of 6.67% while the S&P 500 returned 24%. Moreover, a 2020 study conducted by S&P Dow Jones Indices compared actively managed funds to the performance of the S&P 500, finding that 89% of fund managers failed to beat the benchmark index.

Danelfin's stock analytics platform aims to improve investors' chances of securing large returns by leveraging AI's ability to sort through massive amounts of data in order to provide retail investors with a technological edge formerly reserved for professionally managed funds. Founded in 2016, the company's mission is to "democratize the use of artificial intelligence to help everyone make better investment decisions," according to its website.

Danelfin intends to do so by using its Explainable Artificial Intelligence, an analytics platform, to provide users with stock and ETF ratings, plus an easy-to-understand AI-generated score that ranges from 1 to 10. The platform uses 600 technical indicators, 150 fundamental indicators and 150 sentiment indicators for every stock and ETF it rates. According to the company, the higher the score assigned by its AI, the higher the probability that an equity will outperform the market over the next 90 trading days.

The stocks AI (and my son) picked for me

Because I'm a disciplined investor, I wasn't about to break the bank on an AI stock-picking experiment solely because it would make for a fun story. Fortunately, most major brokerages now offer fractional shares.

So before the market opened on July 8, I chose two stocks that Danelfin's free AI model recommended — big-box retailer Costco Wholesale Corporation (COST) and legacy carrier United Airlines Holding Inc. (UAL) — both of which received AI scores of 10 at the time. Then, after my son became disinterested in the stock randomizer after roughly 30 seconds, I had his pick: car parts retailer AutoZone Inc. (AZO).

That Monday, before the opening bell, I purchased $1 worth of each. COST was trading for $885.67 per share, UAL was trading for $47 per share, and my 5-year-old's pick, AZO, was trading for $2,815 per share. Meanwhile, the S&P 500 began the day at $5,567.19 before hitting a then-record high.

After 30 trading days

Things got off to a rocky start. In the lead-up to this analysis, the S&P 500 had posted a year-to-date return of 17.38%. But summertime volatility suddenly gripped the market, spiking to its highest levels since October 2023 at the tail end of the last bear market.

After the first 30 trading days, on Aug. 16, here's where things stood:

• S&P 500: $5,554.25, down 0.23%
• COST: $870.59, down 1.70%
• UAL: $42.07, down 10.48%
• AZO: $3,212.87, up 14.13%

The S&P 500 was outperforming the two AI picks despite both elevated volatility and a significant tech-fueled sell-off that saw it fall by 8.49% between July 16 and Aug. 5. But the index was able to recover nearly all of those losses by Aug. 16, the 30-day mark of the experiment. At this juncture, the S&P 500 and both AI stock picks were in the red, and it was beginning to look like my son — who believes Santa is real and broccoli is evil — was a stock-picking savant.

After 60 trading days

September is notoriously the worst month of the year for stock performance. According to data from the S&P 500, since 1950, the average return for the index in September is approximately -0.5%, which makes it the only month that has consistently posted a loss over that period.

When the month began, it looked like more of the same. From the last trading day of August through Sept. 6, the S&P 500 experienced another sell-off that drove the index down by 4.25%. But things turned around quickly as foresighted investors scooped up the shares that inexperienced and panicked sellers offloaded. By Sept. 19, the S&P 500 was trading higher than it was before the sell-off began.

After 60 trading days, on Sept. 30, here's where things stood:

• S&P 500: $5,762.48, up 3.50%
• COST: $886.52, up 0.09%
• UAL: $57.06, up 21.40%
• AZO: $3,150.04, up 11.90%

At this point, the two AI picks had an average return of 10.74%, with one stock outperforming the S&P 500 by nearly fivefold and the other trailing the market by more than three percentage points. Meanwhile, my 5-year-old found himself in second place, with a return of 11.90% through 60 trading days, beating the market and one of the stocks AI picked.

After 90 trading days

The third and final phase of this experiment began with a strong October for the broad market and ended with a boost from the 2024 presidential election. In October, the S&P 500 set several more all-time highs, and following a brief pullback at the end of the month, stocks rallied through the final day of the 90-trading day window prescribed by Danelfin's AI.

At the conclusion of those 90 trading days, on Nov. 11, here's where everything stood:

• S&P 500: $6,001.35, up 7.79%
• COST: $932.88, up 5.33%
• UAL: $89.43, up 90.27%
• AZO: $3,173.40, up 12.73%

While the market rally allowed the S&P 500 to outperform one of AI's stock picks — Costco — it was no match for United Airline's gain of more than 90%. Danelifin's proprietary model was able to identify the buy-low opportunity in the beleaguered airline before the stock hit its year-to-date low of $37.88 on Aug. 5. By Nov. 11, shares of UAL had reached their highest price since January 2020.

The takeaway

At the time of writing, shares of United Airlines just hit their all-time high, suggesting that AI is in agreement with my grandfather's investment mantra of patience and discipline. While Danelfin's other pick failed to beat the market, the average return for both stocks over the 90-trading day window was 47.8%, or more than five times the return of the S&P 500 during the same period.

So: It worked.

But had I decided to chose just one of Danelfin's recommendations, I could have either outperformed or underperformed the market, which leads me to believe that — for now — AI's claims are perhaps no more authentic than those made by the average Wall Street suit.

Nobody has a crystal ball. The same goes for emerging technology like stock-picking AI platforms. And for novice or passive investors, the best approach remains investing in lower-risk index funds that historically provide reliable returns.

My son has since returned to focusing on computer games and manipulating me into buying him cups of Cookie Monster ice cream (which contains enough blue food coloring to leave him looking like the flavor's namesake Sesame Street character on a regular basis). That said, I'll never forget that during a stretch of 2024, my 5-year-old randomly picked a stock that not only outperformed the market but held its own against AI.

More from Money:

Online Sports Betting Is Gaining Popularity — at the Expense of Traditional Investing

This Is the Most Boring (but Effective) Way to Become a Millionaire

What Investors Can Learn From the Worst-Performing Stocks of the Year