AI Brain on Stock Market Chart

Can AI Predict the Market? Unveiling ChatGPT's Stock Forecasting Power

Kai Mendoza monogram
Kai Mendoza in Business & Economy October 2026 • 4 min read.

"Discover how ChatGPT's surprising ability to analyze news headlines is shaking up traditional stock market predictions."


The financial world is constantly seeking an edge, a way to predict the unpredictable movements of the stock market. For decades, analysts have relied on traditional methods, poring over financial statements, analyzing economic indicators, and scrutinizing market trends. But what if a new player entered the game, one with the power to process vast amounts of information and identify patterns invisible to the human eye? Enter ChatGPT, an AI language model with a surprising aptitude for stock market forecasting.

A recent study from the University of Florida has unveiled the potential of large language models (LLMs) like ChatGPT to predict stock price movements using a novel source of information: news headlines. The research challenges conventional wisdom, suggesting that AI can not only interpret complex financial data but also anticipate market reactions with impressive accuracy.

This article dives deep into the findings of this study, exploring how ChatGPT achieves its predictive capabilities, the implications for investors and financial professionals, and the potential for AI to revolutionize the future of stock market analysis.

AI Search Multiple angles on this topic

ChatGPT’s Reach and Capabilities

ChatGPT is a generative AI chatbot developed by OpenAI and was originally released on November 30, 2022. It uses generative pre-trained transformers to produce text, speech, and images in response to user prompts. OpenAI describes its conversational format as capable of answering follow-up questions, admitting mistakes, challenging incorrect premises, and rejecting inappropriate requests. The service presents itself as a tool for answering questions, writing, creating images, completing work, and coding.

Capabilities Are Not Forecasting Methods

The listed source describes ChatGPT Images 2.5 as a feature available across desktop, mobile, and web for ChatGPT, ChatGPT Work, and Codex users. It emphasizes templates for formats such as posters and merchandise, along with the ability to add messages, design elements, and styles. This material documents image-generation functionality rather than an accepted method for stock forecasting, so it does not establish how ChatGPT should be evaluated as a market-prediction tool.

A Recent AI Milestone

ChatGPT’s development represents a recent milestone in the public use of conversational artificial intelligence. Its emergence reflects a broader shift toward systems that interact through natural-language prompts and generate responses across multiple formats. However, the available material here does not provide enough evidence to reconstruct the longer history of AI-based market forecasting or identify specific foundational discoveries.

ChatGPT's Unconventional Forecasting Method

AI Brain on Stock Market Chart

The study's central finding is that ChatGPT scores from news headlines can significantly predict daily stock returns, even without direct financial training. This means the AI can assess the sentiment of a news headline and translate that into a likely market reaction. The predictability is stronger among smaller stocks and following negative news, highlighting AI's ability to identify opportunities in less-analyzed corners of the market and amidst pessimistic sentiment.

To understand these results, the researchers developed a theoretical model incorporating information capacity constraints, underreaction, limits-to-arbitrage, and LLMs. This model generates several key predictions, all of which were empirically tested and supported by the data:

  • AI Threshold: There's a critical level of AI capability needed for profitable predictions.
  • Advanced LLMs: Only advanced LLMs can effectively interpret complex information.
  • Market Efficiency: Widespread LLM adoption can enhance market efficiency.
AI Search Multiple angles on this topic

Evidence Remains Limited

The supplied material does not include a research study, dataset, or review evaluating ChatGPT’s ability to predict stock prices. As a result, no firm conclusion about forecasting accuracy can be drawn from these sources alone. Any assessment would require direct evidence, such as a defined testing period, benchmark, and comparison with established forecasting methods.

Limits of the Available Evidence

The available sources do not document specific failed predictions, market losses, or systematic weaknesses in ChatGPT’s stock forecasts. That absence should not be interpreted as proof that the system forecasts reliably. It means only that the supplied material is insufficient to evaluate failure rates or financial risk.

No Direct Benchmark

The supplied sources do not compare ChatGPT’s market predictions with those of analysts, statistical models, or other artificial-intelligence systems. Without a common dataset and evaluation period, relative forecasting performance cannot be established. Claims that ChatGPT is better or worse than another approach would therefore exceed the available evidence.

These results suggest that sophisticated return forecasting is an emerging capability of AI systems. Moreover, these technologies can alter information diffusion and decision-making processes in financial markets. The study introduces an interpretability framework to evaluate LLMs' reasoning, contributing to AI transparency and economic decision-making.

The Future of AI in Financial Markets

The research discussed here is just the beginning. As AI continues to evolve, its role in financial markets will only expand. This means that traders and the wider finance community need to understand the potential and limits to these new AIs. This also means that firms will be focused more and more on AI to help enhance their productivity in the coming years. Whether AI helps to democratize investing remains to be seen, but the technology is likely here to stay.

AI Search Multiple angles on this topic

A Versatile but Unproven Forecaster

The supplied material portrays ChatGPT as a versatile conversational system that can answer questions, write, create images, complete work, and code. It also describes a dialogue format designed to support follow-up questions, error admission, and challenges to incorrect premises. Those capabilities may support research workflows, but the sources do not establish that they translate into dependable stock-market forecasts.

Testing the Forecasting Potential

Future evaluation of ChatGPT as a market tool would need to move beyond general demonstrations of conversational or creative capability. Useful testing could examine predictions against clearly defined historical periods and transparent benchmarks. The supplied sources do not identify a particular forecasting product, dataset, or future research program, so the next steps remain open.

AI Attention and Market Uncertainty

ChatGPT’s release on November 30, 2022, helped place generative AI at the center of public attention and investment, according to the supplied material. That visibility can encourage interest in applying conversational AI to finance. Nevertheless, the sources provided here do not address systemic market risks, data quality, regulation, or the consequences of widespread automated forecasting.

Human Judgment Still Matters

ChatGPT is designed to interact conversationally, answer follow-up questions, and challenge incorrect premises. These features can make it useful as an interactive aid for people exploring information or developing ideas. The available sources do not show that users should replace financial judgment with ChatGPT predictions, particularly because they provide no evidence of forecasting accuracy or investment outcomes.

How headline-based GPT-4 forecasts compare with trading studies

The headline-based result concerns predicting market reactions and later return drift, while the comparator studies evaluate trading rules or strategies against benchmarks. Those are related but not identical tests: a prediction result alone does not establish that a strategy earns returns after costs.[1] [2] [3]

Later work cited in the evidence uses LLMs in distinct ways, including filtering social-media content before forecasting and directly fine-tuning models to map financial news to returns. These approaches provide context, not a direct replication of the headline-scoring result.[4] [5]

GPT-4 headline scores and stock-market reactions

PopulationLarge language models predicting stock market reactions from news headlines[1]
SettingStock market reactions to news headlines[1]
Main outcomePrediction of initial market responses and subsequent drift[1]
Effect sizeApproximately 90% portfolio-day hit rates for the non-tradable initial reaction[1]
Times cited9[1]

A high hit rate for an initial reaction, not a tradable return

  • approximately 90% portfolio-day hit rateFor the initial market reaction, GPT-4 reached approximately 90% portfolio-day hit rates; the paper describes that reaction as non-tradable.[1]Caveat: The measured initial reaction is non-tradable.

Questions the headline evidence leaves open

  • Would the drift signal remain useful outside the small-stock and negative-news cases where the paper says it is especially strong?[1]
  • How much of the forecast advantage depends on model size, and how would it compare under the same test with other model designs or forecasting approaches?[1] [5] [6]

Sources

  1. Can ChatGPT Forecast Stock Price Movements? Return Predictability and Large Language Models (arXiv (Cornell University), 2023)
  2. An Empirical Evaluation of Non-Linear Trading Rules (Studies in Nonlinear Dynamics and Econometrics, 2003)
  3. Employing Machine Learning Algorithms to build Trading Strategies with higher than Risk-Free Returns (International Econometric Review, 2020)
  4. Enhancing the Accuracy of Nikkei 225 Forecasting Via LLM-Based Filtering of Noisy Social Media Data (International Conference on Information and Computer Technologies (ICICT), 2026)
  5. aclanthology.org (aclanthology.org)
  6. Sentiment trading with large language models - UCL Discovery (discovery.ucl.ac.uk)

About this Article -

Written with AI assistance from published research, and reviewed by the Mystum team. See our About page for more information.

This article is based on research published under:

DOI-LINK: https://doi.org/10.48550/arXiv.2304.07619,

Title: Can Chatgpt Forecast Stock Price Movements? Return Predictability And Large Language Models

Subject: q-fin.st cs.cl

Authors: Alejandro Lopez-Lira, Yuehua Tang

Published: 15-04-2023

Everything You Need To Know

1

How does ChatGPT predict stock price movements?

ChatGPT predicts stock price movements by analyzing news headlines. The AI assesses the sentiment of a news headline and translates that into a likely market reaction. The study from the University of Florida found that ChatGPT scores from news headlines can significantly predict daily stock returns, showcasing its ability to interpret complex information without needing direct financial training. This approach allows ChatGPT to identify patterns and anticipate market reactions, particularly in smaller stocks and in response to negative news, where the predictability is stronger.

2

What is the significance of the AI Threshold in the context of stock forecasting?

The AI Threshold represents a critical level of AI capability needed for profitable predictions. The study's theoretical model incorporating information capacity constraints, underreaction, limits-to-arbitrage, and LLMs, highlights that not all AI systems are created equal. Only advanced LLMs, like ChatGPT, can effectively interpret complex information and translate it into actionable insights for stock forecasting. This implies that as AI technology advances, the ability to cross this threshold becomes increasingly important for accurate market predictions. Reaching this threshold is vital for AI systems to achieve significant predictive accuracy and influence in financial markets.

3

How might the rise of AI like ChatGPT impact investment strategies?

The rise of AI like ChatGPT has the potential to reshape investment strategies by providing a new tool for market analysis and prediction. Investors and financial professionals may need to adapt their strategies to incorporate AI-driven insights. This could mean leveraging AI to analyze vast amounts of data, identify market trends, and make more informed investment decisions. AI can also enhance market efficiency by altering information diffusion and decision-making processes. Investors should be aware of the potential and limits of these new AIs to make informed decisions. Whether AI helps to democratize investing remains to be seen, but the technology is likely here to stay.

4

What are the key findings from the University of Florida study regarding ChatGPT's capabilities?

The University of Florida study revealed that ChatGPT can predict daily stock returns by analyzing news headlines. The study showed that ChatGPT's scores from news headlines significantly predict stock returns, even without financial training. The AI demonstrates strength in assessing the sentiment of news and anticipating market reactions, especially among smaller stocks and in response to negative news. The researchers also developed a theoretical model that incorporates concepts like AI Threshold, Advanced LLMs, and Market Efficiency, which were all supported by empirical data. These findings highlight the potential of sophisticated return forecasting as an emerging capability of AI systems, potentially changing information diffusion and decision-making in financial markets.

5

What are the potential implications of widespread LLM adoption in financial markets?

Widespread adoption of Large Language Models (LLMs) like ChatGPT can enhance market efficiency. As more market participants utilize advanced AI, the ability to analyze and interpret information becomes more sophisticated, potentially leading to more accurate price discovery and reduced market inefficiencies. Sophisticated return forecasting is an emerging capability of AI systems. Increased adoption of LLMs can lead to faster information diffusion and potentially alter how decisions are made in financial markets. This could, in turn, impact how investors and financial professionals operate, creating a more dynamic and potentially more efficient market environment, although not necessarily one that is easier to navigate.

Newsletter Subscribe

Subscribe to get the latest articles and insights directly in your inbox.