You may not be an Ivy League research analyst with access to a $25,000-a-year Bloomberg terminal…

But what if you had something better – an AI stock picker that could outperform the pros using nothing but headlines, data, and a little reasoning?

That’s the question Alejandro Lopez-Lira, a finance professor at the University of Florida, has spent the last few years exploring.

With tools like ChatGPT, Grok, and DeepSeek at his fingertips, Lopez-Lira wanted to know: Can artificial intelligence actually pick winning stocks?

Turns out, it can.

The First Experiment: Backtesting GPT on News Headlines

Lopez-Lira began his AI investing journey in late 2022, shortly after OpenAI launched ChatGPT.

He ran a study using ChatGPT to interpret over 134,000 stock-related headlines.

Each day, the AI would label each headline as good, bad, or neutral for a stock.

If it was good, it simulated a buy. If bad, it simulated a short. Neutral? No trade.

To prevent hindsight bias, he only used headlines from after September 2021, the cutoff date for ChatGPT’s training data.

This forced the model to reason – rather than “remember” – how a stock might react.

The results were staggering: a simulated daily return of 0.38%, which compounded to more than 650% over 26 months.

A big chunk of the performance came from shorting stocks with negative headlines.

Was it perfect? No.

The backtest didn’t account for real-world frictions like taxes, short interest fees, or trading costs.

But it was impressive enough for Lopez-Lira to take the idea out of the lab and into the market.

Round Two: Investing Real Money With AI

In 2023, Lopez-Lira teamed up with the Autopilot app to start live-testing AI-generated portfolios.

He used ChatGPT to pick stocks using real-time financial data, macro headlines, and pricing information.

He later added Grok and DeepSeek to the experiment.

Each model was given the same data and asked to build a 15-position portfolio.

The AIs were allowed to decide their own mix of assets and even allocate weights.

The only limits? No leverage, no short positions, and no derivative-based ETFs.

Here’s what each AI picked for the volatile April 1, 2025 to May 5, 2025 cycle:

OpenAI’s GPT-4 (o1 pro) Portfolio Picks

Strategy: Broad market exposure, tech giants, and stable dividend payers.

Grok 3 Portfolio Picks

Strategy: Defensive and yield-focused with heavy Treasury exposure. Grok leaned into safety ahead of potential economic softness.

DeepSeek R1 Portfolio Picks

Strategy: Blue-chip stocks with defensive traits. A “stay in the game” strategy in volatile conditions.

So, Should You Trust AI With Your Money?

Lopez-Lira is the first to admit that these tools still need oversight.

They don’t access real-time data by default, and their decisions are only as good as the information they’re fed.

But the early results are promising:

• Lopez-Lira began managing ChatGPT’s portfolio in September 2023. The returns are 43.5% from September 2023 to May 30, 2025 while the S&P 500 had a total return of 34.7% over the same period.

• Grok’s portfolio returned 2.3% since its inception on Feb. 11 of this year through May 30. The S&P 500 had a total return that was down 2.2% over the same period.

• DeepSeek was down 0.25% since its inception on Feb. 3 through May 30. The S&P 500 had a negative total return of 0.93% for the same period.

You can invest alongside these portfolios yourself on the Autopilot app.

[Note: We do not have any affiliation with Autopilot.]

As AI keeps improving, it won’t be long before investors without fancy credentials or six-figure salaries can compete with Wall Street.

No PhD required – just a smart prompt.