Can AI Help You Beat the Stock Market? The Answer May Surprise You

Artificial intelligence became a household investing topic after ChatGPT debuted in November 2022. But Wall Street had been teaching machines to analyze markets for decades. Quantitative funds have used statistical and mathematical computer models since the 1970s, while asset managers have increasingly turned to machine learning to process data humans cannot digest at comparable speed. 

The investing thesis is straightforward: AI can give investors a powerful analytical advantage, but owning an AI-powered strategy does not automatically mean beating the market.

AI Stock Picking Predates ChatGPT

Retail investors quickly tried using ChatGPT and other large language models to pick stocks. The early results were often underwhelming. Hallucinations — when an AI confidently produces incorrect information — could turn an apparently sophisticated investment thesis into fiction.

Today’s leading models are far better at reasoning, retrieving information, and reducing hallucinations, although none is error-free. That matters because professional investors have been applying more specialized forms of AI for years.

Amplify Investments launched the Amplify AI Powered Equity ETF (NYSEARCA:AIEQ) on Oct. 17, 2017, making it an early public test of machine-driven stock selection. Its strategy uses IBM‘s (NYSE:IBM | IBM Price Prediction) Watson through EquBot to analyze financial statements, news, sentiment, macroeconomic information, and other data.

Then came fintech-native managers. Qraft Technologies launched Qraft AI-Enhanced U.S. Large Cap Momentum ETF (NYSEARCA:AMOM) and Qraft AI-Enhanced U.S. Large Cap ETF in May 2019, using machine learning to adjust factor exposure and select stocks. Following generative AI’s explosion in 2022, institutional adoption accelerated as established asset managers began deploying proprietary deep-learning models for stock selection.

Nearly a Decade Of Results Says Plenty

Here’s where things get interesting. AIEQ has been operating for nearly nine years, giving investors a meaningful window into whether a machine can actually deliver superior returns.

Since inception, AIEQ has gained 100.2%. That sounds pretty good until you put it next to the benchmark. Over the same period, the S&P 500 has gained roughly 199.7% — almost twice as much.

That’s an important reality check. The world’s first fully AI-driven ETF didn’t implode, but it also didn’t demonstrate that artificial intelligence could consistently outsmart a simple index fund.

The newer generation has produced a more encouraging snapshot. Based on current YTD and one-year returns:

AMOM is the standout, beating the S&P 500 by more than 7 percentage points year-to-date and roughly 5 percentage points over one year. But investors should resist extrapolating that short-term advantage into a permanent AI edge. Several of these funds are too young to have established a meaningful long-term record.

The broader lesson is that AI’s investment record remains mixed. Early predictive systems had particular difficulty with events outside their historical training data, including the 2020 pandemic and the inflation shock that followed. A model can process millions of observations and still struggle when the future produces something the past never contained.

The True AI ETF Universe Is Smaller Than It Looks

Search a brokerage app for “AI ETFs” and you’ll typically encounter funds such as Global X Artificial Intelligence & Technology ETF (NASDAQ:AIQ), Global X Robotics and Artificial Intelligence ETF (NASDAQ:BOTZ), or Roundhill Generative AI & Technology ETF (NYSEARCA:CHAT). Those are AI-themed ETFs. They invest in companies benefiting from artificial intelligence; they do not necessarily use AI to select their holdings.

The true AI stock-picking universe is much smaller:

WisdomTree’s AIVL uses a proprietary quantitative AI model to identify and overweight undervalued U.S. equities.

There is also a practical problem. True AI-managed funds generally charge much more than passive index funds. A 0.75% annual expense ratio is 25 times the 0.03% fee available from some S&P 500 ETFs. Smaller AI funds can also struggle to attract enough assets to remain viable, as the liquidation of QRFT demonstrated.

In short, investors shouldn’t confuse increasingly sophisticated technology with proven investment alpha. AI has become dramatically better since 2022, and professional AI investing has a much longer history than ChatGPT. But the evidence still doesn’t show that handing stock selection to a machine reliably beats a low-cost S&P 500 index fund.

Key Takeaway

AI may eventually become a better stock picker than humans, but investors don’t need to wait for that verdict. The current evidence suggests a more measured approach: use AI as a powerful research tool, and evaluate AI-managed ETFs the same way you would any active manager — by demanding a long track record, competitive fees, adequate liquidity, and performance that survives comparison with the S&P 500.

AMOM’s recent numbers show that an AI strategy can beat the market. AIEQ’s nearly nine-year record shows that doing it consistently is another matter entirely.

Contact [email protected] for any questions or corrections.