Can AI Predict the Stock Market? How ChatGPT Outshines DeepSeek

Introduction

Imagine if an AI could read financial news and predict stock market movements. Wouldn’t that be a game-changer for investors? Well, that’s exactly what researchers set out to explore in a recent study comparing ChatGPT and DeepSeek, two powerful AI models, to see how well they could predict stock market trends and broader economic conditions based on news from the Wall Street Journal.

The study found that ChatGPT successfully predicted future market returns, while DeepSeek struggled to do the same. But why is this the case? And what does this mean for financial forecasting and investor behavior? Let’s break it down in simple terms.

How Can AI Predict the Stock Market?

For decades, investors and economists have looked for ways to predict stock market trends. Traditional methods rely on financial indicators such as earnings reports, interest rates, and economic data. But recently, AI models—trained on vast amounts of text—have stepped into the game, attempting to extract insights from news articles and financial reports.

This study leveraged two AI models:

1. ChatGPT – A widely known AI model designed to process and analyze human language, optimized for English.
2. DeepSeek – Another AI model with strong multilingual capabilities, though trained more extensively in Chinese.

By feeding these models data from Wall Street Journal headlines and alerts from 1996 to 2022, the research tested whether they could identify good and bad news—and if those insights could predict the stock market’s movement.

What the Study Found

1. ChatGPT Can Predict Stock Market Trends—DeepSeek Can’t

ChatGPT successfully identified important financial signals from news headlines and used that information to predict how the stock market would behave in the next six months.

• Headlines that contained “good news” were positively correlated with future market returns.
• Interestingly, bad news didn’t predict future trends. Instead, the stock market reacted almost immediately to bad news, making it ineffective for forecasting.

In contrast, DeepSeek captured the current market sentiment but failed to predict how the market would behave in the future. Since DeepSeek is trained more in Chinese and less in English, it may have struggled with the nuances of analyzing Wall Street Journal data.

2. Good News Takes Longer to Influence the Market than Bad News

The study confirmed something many investors have long suspected: markets respond slower to good news than to bad news. Investors tend to immediately react to bad news (like economic downturns or company scandals), while good news takes longer to fully shift investor confidence and stock prices.

This delayed reaction to good news allowed ChatGPT to identify profitable investment opportunities that human analysts might overlook.

3. Why Did ChatGPT Succeed? The Power of Large Language Models

The study compared ChatGPT with other methods, including:
– The traditional ‘word list’ method, which classifies words as positive or negative based on predefined dictionaries (e.g., sentiment analysis).
– Smaller AI models like BERT, which lacked the ability to interpret words based on context.

Neither of these approaches worked as well as ChatGPT. Why?

Because large AI models like ChatGPT don’t just classify words—they understand context. For example, in standard sentiment analysis, the word “growth” is always seen as positive. But ChatGPT can recognize when “growth” refers to a company’s expansion and when it’s referring to “growth in unemployment” (which is bad news).

Real-World Implications: How Could This Change Investing?

The ability of AI to predict market trends has huge implications for investors, financial analysts, and policymakers.

1. Investment Opportunities

If AI can detect undervalued stocks based on delayed reactions to good news, hedge funds and large investors could develop AI-driven investment strategies to capitalize on these gaps.

2. Market Efficiency Debate

Traditional finance holds that markets are “efficient”—meaning that all available information is immediately included in stock prices. But this study suggests not all information is absorbed instantly, particularly good news. This raises questions about whether markets are truly as efficient as economists believe.

3. AI-Powered Financial Analysis

This study shows that ChatGPT outperforms traditional financial news analysis. This could revolutionize how news sentiment analysis is conducted in the financial sector and may lead to more AI-powered financial decision-making.

Can You Use ChatGPT for Your Own Market Predictions?

The idea of using AI to predict market trends is exciting, but can YOU use ChatGPT to do the same thing?

The answer is: Yes, but with limitations.

What you can do:
✅ Use ChatGPT to summarize and interpret financial news faster.
✅ Ask ChatGPT to classify financial headlines as “positive” or “negative” to gauge market sentiment.
✅ Experiment with its ability to detect financial trends.

What you can’t do:
❌ Rely on ChatGPT as a financial oracle—it still has flaws.
❌ Use it for real-time price predictions (it lacks up-to-date market data).
❌ Assume it will always beat traditional investment strategies.

ChatGPT is a great tool for improving your financial literacy and investment research—but it’s always good to combine AI insights with traditional financial analysis.

Key Takeaways

✅ ChatGPT can read financial news and predict stock market trends—DeepSeek struggles with this task.
✅ Good news tends to influence stock prices more slowly than bad news, creating opportunities for AI-driven predictions.
✅ Traditional methods (like sentiment analysis and smaller AI models) fail where ChatGPT succeeds, as ChatGPT understands words in context.
✅ This AI-driven market forecasting raises questions about how efficiently financial markets process news.
✅ While ChatGPT is promising for financial research, it shouldn’t be treated as a foolproof stock-picking tool.

Final Thoughts

This study marks an important step toward understanding how AI can complement and even surpass human financial analysis. While we’re not at the point where we can trust AI to make investment decisions for us entirely, we are seeing the dawn of AI-assisted investing.

As AI models continue to improve, they will likely play a bigger role in financial decision-making, investment forecasting, and even macroeconomic policy analysis. For investors, keeping an eye on how AI is used in finance could unlock new opportunities and strategies that were previously impossible.

And who knows? Maybe in the near future, your AI assistant will be able to pick your next winning investment! 🚀

If you are looking to improve your prompting skills and haven’t already, check out our free Advanced Prompt Engineering course.

This blog post is based on the research article “ChatGPT and Deepseek: Can They Predict the Stock Market and Macroeconomy?” by Authors: Jian Chen, Guohao Tang, Guofu Zhou, Wu Zhu. You can find the original article here.