In November 2022, a new chatbot landed on the internet and pulled in a million users within a week. ChatGPT could draft emails, write code, and answer questions in fluent prose, and within two months it had reached roughly 100 million users. Its arrival set off conversations in nearly every profession about what artificial intelligence might mean for work. But a quieter question lingered around financial markets: did the buzz around this new tool change how investors actually felt about buying and selling?
A team of researchers set out to investigate exactly that. Writing in the European Journal of Sustainable Development, they looked at whether the launch of ChatGPT coincided with a measurable shift in investor mood across both the stock market and the cryptocurrency market. Their analysis offers evidence that sentiment in both markets rose after the chatbot appeared, though the tool explains only a small slice of the overall picture.
The question behind the study
Plenty of earlier research had already examined whether ChatGPT could help people make money. Some studies tested whether it could pick stocks, build portfolios, or forecast returns from news headlines. The authors of this paper, led by Adel Barguellil of the Ecole Supérieure de Commerce de Tunis at Manouba University in Tunisia, working with colleagues at King Abdulaziz University in Saudi Arabia, point to a different gap. Almost no one had asked whether ChatGPT influenced investor sentiment itself, meaning the collective emotional temperature of the market rather than any specific trade.
That distinction matters because emotions tend to drive financial decisions. When investors feel confident, they buy more and push prices up. When they feel anxious, they pull back. The researchers were curious whether a high-profile AI tool could nudge that mood, and whether such a nudge might add to confidence, or instead feed herd behavior and speculative bubbles, especially in a volatile arena like crypto.
Two thermometers for market mood
To track sentiment, the team leaned on two widely used gauges. The first is the Crypto Fear and Greed Index, which runs on a scale from 0 to 100. A low number means investors are fearful, while a high number means they are greedy and chasing gains. The second is the American Association of Individual Investors Sentiment Index, which surveys ordinary investors on whether they expect the stock market to rise (bullish), fall (bearish), or stay flat (neutral) over the next six months.
The researchers combined those survey percentages into a single weighted score, adding the bullish and neutral shares and subtracting the bearish share. They gathered weekly readings for both indices stretching from January 1, 2020, through December 1, 2024, giving them 255 weekly observations to work with.
Drawing a line at the launch date
To test whether ChatGPT’s release made a difference, the team used a method called interrupted time series analysis. The idea is straightforward. You track a measurement over a long stretch of time, mark the moment an event happened, and then check whether the pattern before that moment looks different from the pattern after it. The event here was ChatGPT’s public launch in November 2022. The researchers built a statistical model that flagged every week as either “before” or “after” the launch and looked for shifts in the level and direction of each sentiment index.
For the Crypto Fear and Greed Index, the model found that the launch was linked to an increase of about 11 points, a result that was statistically significant. For the stock market’s AAII index, the jump was even larger at roughly 12.6 points, also statistically significant. In plain terms, both gauges sat higher after ChatGPT arrived than the pre-launch trend would have predicted.
At the same time, the researchers are candid about how much their model actually explains. For the crypto index, the launch accounted for only about 6.5 percent of the variation in sentiment. For the stock index, the figure was about 10.2 percent. The bulk of the mood swings in these markets came from other forces entirely. The authors note that “ChatGPT’s influence, while statistically significant, operates within a complex behavioral ecosystem,” pointing to macroeconomic conditions, media coverage, and geopolitical events as factors likely interacting with any AI-driven optimism.
Imagining a world without the chatbot
As a second check, the team turned to a forecasting tool called Facebook Prophet. They used it to build a hypothetical scenario: what would each sentiment index have looked like if ChatGPT had never been released? The model produced a predicted trajectory based on the patterns that existed beforehand, and the researchers then compared that prediction against what really happened.
For both indices, the actual readings after the launch ran above the predicted ones. The crypto index climbed sharply past its forecast, while the stock sentiment index showed both an upward drift and more volatility than the model expected, with a gap that was especially visible around the middle of 2023. The authors interpret these gaps as evidence that ChatGPT played what they call a “catalytic role” in behavioral changes across both markets.
Mapping the research field
Before running their main analysis, the team also conducted a bibliometric review, which is essentially a study of the existing studies. They tallied which countries, authors, and journals were most active in research connecting ChatGPT and finance. China, the United States, and India produced the most publications, and journals such as Expert Systems with Applications and Finance Research Letters emerged as central outlets. This exercise helped the authors situate their own work and identify investor sentiment as an underexplored thread.
What to make of the findings, and what not to
The study fits within a larger argument the authors are making about sustainable finance. They contend that tools like ChatGPT should be judged not only by short-term profits but by their contribution to transparency, resilience, and inclusiveness in financial systems. They frame generative AI as having a dual nature: a disruptive force capable of reshaping markets, and a possible driver of more responsible financial practices, provided that issues of access, governance, and ethical oversight are addressed.
A few caveats deserve attention. This is observational research, not an experiment, so it cannot prove that ChatGPT caused the rise in sentiment. The early months of 2023 were eventful for markets in many ways, and the modest variance the model explains is itself a reminder that countless other factors were at play. The interrupted time series approach can show that something changed around a date, but it cannot fully separate the chatbot’s effect from everything else happening at the same time.
Still, the researchers suggest some takeaways. They argue that policymakers should understand how AI tools can sway market mood when designing rules meant to limit volatility. They also suggest that investors would benefit from awareness of how ChatGPT and similar technologies may shape their own decision-making, sometimes amplifying both optimism and uncertainty. As the authors put it, future work should assess “whether reliance on AI strengthens financial resilience or introduces new systemic vulnerabilities.”
