Investors have a lot of information to sort through when deciding where to put their money: earnings reports, financial statements, analyst forecasts, company news, and market shifts.

Institutional investors — think mutual funds, pension funds, and hedge funds — have long had an advantage in making sense of it all. They have professional analysts, specialized tools and data, and, perhaps most importantly, people whose full-time jobs are to understand what the numbers mean.

Individual investors have had to make many of the same decisions without those same resources.

Enter generative AI.

Individual investors now have access to a tool that can summarize an earnings report, explain why a stock moved, compare companies, and help make sense of dense financial information in seconds.

For Foster accounting professor Elizabeth Blankespoor, that raised an obvious question: What are individual investors, also called retail investors, doing with this new capability, and could it help level the playing field?

“This tool could be amazing for them,” Blankespoor said. “It could really help reduce their disadvantage, or maybe not.”

In new research published in the Journal of Accounting and Economics, Blankespoor and her coauthors found that individual investors used generative AI for a wide range of tasks, particularly to interpret financial information and understand market movements. They also found that investors with more finance and accounting knowledge tended to adopt the technology more readily, use it for more complex tasks, and draw on a broader range of information sources.

In other words, AI can make financial information easier to access and process. But it does not make expertise irrelevant.

A window into real-world AI use

To understand how investors were using AI, Blankespoor and her colleagues drew on two complementary sources of information.

One was a large dataset of questions investors asked Alpha, a generative AI chatbot built into the investing platform Public. The researchers analyzed more than 400,000 investor queries collected between January 2024 and April 2025, giving them a direct look at how people were using GenAI in an investing context.

“We didn’t have to guess or say, ‘Imagine that you’re thinking about investing,’ ” Blankespoor said. “No, these were investors. They’re using it.”

The query data showed what investors asked, but the researchers could not know why someone chose to use AI, why someone else did not, whether users found the tool helpful, or how much financial knowledge they brought to the task.

To fill in those gaps, the researchers surveyed more than 2,000 individual investors about their financial backgrounds, their reasons for using or avoiding GenAI, the tasks they used it for, and their concerns about the technology.

The two sources answered different parts of the same question, and the findings lined up in important ways. The tasks and information sources investors said they used AI for in the survey closely matched what the researchers saw in the real-world queries.

Nearly half of the investors surveyed in April 2024 said they were already using GenAI to process financial information or inform investment decisions. Blankespoor suspects that share would be substantially higher today, although the researchers have not yet updated the survey.

New tools for researchers, too

AI is not only the subject of Blankespoor’s research. It is also widening the scope of what’s possible in accounting research.

Analyzing hundreds of thousands of investor questions by hand would have been extraordinarily difficult. Instead, the researchers manually classified a subset of questions, trained a machine-learning model to classify the rest, checked its accuracy, and then applied the approach across the larger dataset.

Classification algorithms have existed for years, Blankespoor notes. But newer AI tools can work across large amounts of text with less manual training, while still requiring researchers to validate the results.

“It opens up questions that are more difficult and require more nuanced classification of text than we could do in the past,” Blankespoor said.

How ai in investing changes with financial expertise

Investors used GenAI to help interpret financial information and market movements, screen potential investments, and streamline more complicated research.

How they used those capabilities differed with their level of financial knowledge.

Investors with more finance and accounting experience asked more complicated questions and used AI for more demanding tasks. Those with less background tended to ask more straightforward questions: Can you explain this to me? What should I do? What does this market movement mean?

The researchers also saw the way some investors used AI shift over time. Among people who returned to the brokerage’s AI tool repeatedly, questions shifted somewhat away from broad stock screening and high-level company assessments and toward more detailed monitoring and interpretation of company news.

Taken together, those patterns suggest that getting more from AI depends in part on knowing enough to use it well.

“If you don’t have a base knowledge, it’s hard to know what questions to ask,” Blankespoor said. “And it’s hard to know what’s an important thing to chase down and explore further versus not. What’s something to push back on?”

Building the judgment to use AI well

Blankespoor is seeing that same relationship between AI and expertise play out in her financial accounting classroom.

In one assignment, she asks students to use AI to find return on assets, a common financial metric, for two companies. With one of the companies, the task was relatively straightforward. With the other, the AI returned two possible numbers, and the students were unsure which one to use.

The complication involved something they had not yet learned in class. Blankespoor could look at the answer and understand why the two numbers appeared. Her students did not yet have the accounting knowledge to do the same.

That is precisely the challenge she wants them to recognize.

As an accounting student or a new investor, “You can’t just take whatever number AI gives,” Blankespoor said. “You’re going to have to make a judgment call. And you need to know accounting if you’re going to make that judgment call.”

That judgment becomes especially important when an AI-generated answer is inaccurate.

“If you don’t really know very much, it’s hard to push back on accuracy,” Blankespoor said. “Whereas if you already know a lot more, then you can.”

But Blankespoor also sees another way students can respond when they reach the limits of what they know: They can use AI to learn.

A student who does not understand why the tool produced two possible answers can keep asking questions, using the AI to explain the underlying accounting concept until the distinction makes sense.

Blankespoor sees a broader lesson in both the research and the classroom: AI can help people work more efficiently and effectively, but it can also help them build the knowledge they need to use it better.

Learn more about AI at Foster.

Read the full research paper: Blankespoor, Elizabeth and Croom, Joe and Grant, Stephanie M., Generative AI and Investor Processing of Financial Information (June 17, 2026). Forthcoming, Journal of Accounting and Economics, Volume 82, No 2, November 2026.