How Wharton Is Preparing Students for an AI-Driven Financial Future
AI is not only changing how financial markets move; it’s changing who moves them. Decisions once made by analysts and traders are increasingly delegated to learning algorithms that can operate autonomously and process data on a scale no human can match. Firms hiring today’s graduates expect them not only to understand this new market order, but to compete in it and help define the next frontier of financial practice.
That transformation is at the heart of AI and Data Science in Finance, a course taught at undergraduate and graduate levels by Winston (Wei) Dou, associate professor of finance at Wharton and faculty director of the Wharton AI in Finance Lab. Dou’s research examines strategic interactions in financial markets, with a growing focus on what happens when algorithms themselves become market participants. The course brings those questions, and the frontier of his research, directly into the classroom.
“This course is not only about coding or financial theory,” Dou says. “It’s an opportunity for students to learn a new way of thinking — how to ask deep questions in the financial world, and how to address those questions from a modern data science perspective.”
Beyond the Prompt: Teaching a Way of Thinking, Not Just a Tool
Dou is emphatic that his course is not a tour of whichever AI product happens to be popular that semester. Students move methodically through the field, from linear and nonlinear prediction models to textual analysis and LLMs, then to reinforcement learning for autonomous decision-making. They build each tool from first principles rather than starting from a chat window. “We know the most popular way people use AI nowadays is to write prompts and get answers,” Dou says. His students go several layers deeper. “The key point of this class is to teach students a way of thinking: how to think about and solve real financial problems using the ideas and architecture of AI. That way of thinking will last much longer” than any single tool’s shelf life.
LEARNING OBJECTIVES
- Understand the fundamental ideas, algorithms, and architectures behind AI, machine learning, and data science in finance.
- Develop skills to formulate financial challenges as data-driven problems and apply appropriate AI methods to areas such as asset pricing, trading, and risk management.
- Critically evaluate the potential and limitations of AI in financial markets, including model failure, data bias, and risks created by autonomous systems.
That grounding matters because, as Dou readily acknowledges, AI in finance is not a solved problem. “I’m very excited about AI and very optimistic about it,” he says, “but we need to embrace AI with caution. AI can make mistakes, be hard to understand, and be trained on biased data. Most important, when AI systems interact with one another in financial markets, they can generate unpredictable outcomes.”
That last point, what happens when many AI systems trade at once, is one of the course’s most distinctive threads. It is also drawn directly from Dou’s research. An individual firm that delegates its investment decisions to autonomous AI agents that learn from experience may perform better. Scale that up across many firms deploying similar AI agents, however, and the picture becomes more complicated.
“That type of better trading performance is not necessarily because the machines are smarter than humans,” Dou explains. “It can instead come from the way machines interact with one another, forming a kind of coordination and earning trading profits at the expense of other investors.”
In other words, AI trading systems trained independently to maximize their own profits can, through repeated market interaction, end up creating the very kinds of market inefficiencies that AI is often expected to eliminate.
Real Data, Real Stakes
Those ideas are tested against real market data in the course’s capstone project. Students choose from a menu of real-world financial problems, including evaluating investment performance, forecasting prices, and building autonomous trading AI agents benchmarked against idealized “Oracle” performance. They then decide for themselves which data to use, and which methods best fit the problem.
Antonis Kriezis, SEAS ’26, who pursued an accelerated master’s degree in scientific computing, used the project to examine a long-running debate in finance: whether mutual fund managers possess genuine skill or whether their returns can be explained by simple, replicable investment strategies. His team found that many funds posting strong returns were riding popular investment themes, such as technology or high-growth stocks, rather than demonstrating stock-picking skill. Investors could often obtain similar exposure more cheaply through a low-cost index fund, without paying active-management fees.
“People get confused because they correlate the success of a mutual fund manager with their returns,” he says. “In reality, it’s more nuanced than that.”
Although the course was not required for his degree, Kriezis says AI and Data Science in Finance filled a gap he had been unable to fill elsewhere at Penn. “Everything we learned was tied directly to a specific application, from asset pricing and credit risk to portfolio construction,” he says. He credits Dou with keeping that connection explicit throughout the semester. “It was obvious he really cared about both his students and his research.”
“The key point of this class is to teach students a way of thinking: how to think about and solve real financial problems using the ideas and architecture of AI.”
Winston (Wei) Dou, Instructor, AI and Data Science in Finance
From Classroom to Career
Anika Bastin, W’27, a senior concentrating in statistics and finance, and a student in the newly launched Dr. Bruce I. Jacobs Master of Science in Quantitative Finance program, came to the course well-versed in both machine learning and finance, but she had always studied them separately. “I’ve done classes in ML, I’ve done classes in finance separately, but none of them really bridged the gap,” she says. “This was a cool opportunity to bring together those two domains.”
What stood out to her was how the course handled the particular messiness of financial data. “Finance is a whole different space. There’s so much noise in the data, so many patterns you don’t get in other types of data,” Bastin says. “Having a course study that specifically was groundbreaking compared to literally every other course I’ve taken.”
Bastin, who has worked in equity research and was classmates with Kriezis in the course’s undergraduate offering, is now applying those same ideas to her senior thesis, advised by Dou. She is examining whether LLMs have changed how equity analysts write research reports and issue investment recommendations. Her preliminary work suggests that AI tools may be helping analysts extract information more efficiently, though she cautions that the research remains at an early stage.
Both students describe an AI-augmented workforce as more of an adaptation than a replacement. “AI in my life so far is a very good add-on,” Bastin says. “I still need to be able to perform on the spot, but AI is like a very good study tool to help me think more out of the box.”
Kriezis, who joined a hedge fund in New York as a quantitative researcher after graduation, sees the same dynamic playing out in industry. “The problem isn’t access to data anymore, it’s how to make sense of it,” he says. “The people who are going to excel are the people who know how to convert data into actionable decisions using AI as a tool for it.”
Looking Ahead: A New Lab for AI and Finance at Wharton
Dou’s ambitions extend beyond the classroom. He is spearheading the Wharton AI in Finance Lab (WAFL), a new initiative that will carry the course’s core questions into a broader program of research and industry collaboration.
“The intersection of AI and finance will be one of the most important frontiers in both practice and research,” Dou says. WAFL will bring together faculty, students, researchers and practitioners to exchange ideas, build shared data infrastructure and support student projects, while remaining closely connected to the challenges facing practitioners on Wall Street. “The goal is for Wharton to be not only a place where students develop new skills, but also an important force leading this transformation in finance.”
It’s a fitting next chapter for a course built on the idea that the tools will keep changing — but the way of thinking behind them doesn’t have to start over each time. As Dou puts it: “This course will plant a seed in their mind, which will grow after their further studies, research, or practice on the path of their career in this area.”