Teaching Through a Technological Revolution

For over a decade, Song Ma and his students have collected and cleaned data from the U.S. Patent and Trademark Office for his research on innovation and entrepreneurship, tabulating records on nearly seven million patents granted over several decades. The work used to take about a week each year as it involves mapping patents to millions of different firms, institutions, and inventors.

Not anymore. Ma, a professor of finance and entrepreneurship, recently created an artificial intelligence agent that will complete the task automatically, correcting irregularities, with Ma and his team spending just minutes reviewing anomalies before posting to the public and sharing with other researchers.

“In the past, it could take months for a new researcher, like a PhD student, to understand the data and the algorithms well enough to work with them effectively,” Ma says. “Now, the process is much more sustainable because the knowledge is embedded in the documentations and code themselves, and AI can help onboard any new researcher. This applies to many other research tasks that my SOM faculty colleagues do on a daily basis, from data collection, data cleaning, exploration of research ideas, clean documentations, among others.”

Ma is one of many SOM faculty exploring the possibilities and pitfalls of AI. He believes that the technology is transforming the world, and that scholars at leading institutions like Yale SOM must play an important role in shaping its impact on business and society.

Ma regularly joins his SOM faculty colleagues for a seminar series on computing—which lately has focused on AI. In the seminar, co-organized by Kyle Jensen, Shanna and Eric Bass ’05 Director of Entrepreneurial Programs, associate dean, and professor in the practice of entrepreneurship; Paul Goldsmith-Pinkham, associate professor of finance; and Kevin Williams, professor of economics, faculty discuss how AI can be used to accelerate faculty research and how to integrate AI into the classroom.

Recognizing the importance of the subject to every business student, the senior faculty voted in May 2026 to create a new MBA core course that will introduce students to AI tools and to expand consideration of AI throughout the core.

Ma, who will lead a conference on AI in the classroom in the fall of 2026, said the urgency of this work was made clear when he led students on a weeklong trip to Silicon Valley, where they met with alumni working in venture capital and technology. AI, he said, was omnipresent.

“It was mind-blowing,” Ma says. “The students realized: This is coming, you need to be on top of this.”

Creating a New Collaborator

Tong Wang, assistant professor of marketing, teaches the course AI for Business Decisions, which uses AI to predict patterns at scale—such as whether a product is likely to be returned. In programming-intensive courses like this one, she once handed students coding templates. Now she teaches them to direct an agentic AI assistant instead—a change, she says, in what students actually need to master.

“In the past, students had to master two things at once: the high-level reasoning—how to frame a business problem as a model—and the implementation: the technical details underneath it. And the details often crowded out the thinking,” Wang says. “Now AI handles the implementation, so students can put their full attention where their judgment actually matters: what problem to solve, how to structure it, and how to tell whether the answer is right.”

Wang’s pedagogy is grounded in her research. She was applying machine learning to business problems long before LLMs reached the public; in 2019, she was part of the team that won the inaugural FICO Explainable Machine Learning Challenge. That background—building models designed to explain themselves—shapes how she sees the technology today: not as a tool, but as a new kind of collaborator.

But reliable collaboration with AI doesn’t happen automatically. When Wang enlists AI as a “beta reader” on her papers and presentations, she deliberately counters its “sycophancy”—the tendency to flatter users—by instructing it to “criticize this idea” rather than asking what it thinks of one.

“We’ve essentially created a new collaborator,” she says. “But a good collaborator tells you the truth, not what you want to hear. Out of the box, AI wants to please you—so part of the skill now is shaping it into a collaborator you can trust to push back.”

Challenging Students

Paul Goldsmith-Pinkham describes himself as a “power user” of AI agents in his research and teaching. In one of his classes, Goldsmith-Pinkham crafted a contrarian AI agent to discuss case studies with students and force them to defend their ideas rigorously.

But he doesn’t hesitate to explore the pitfalls of the technology and its potential to exacerbate existing inequities. One of his papers, which won the American Finance Association’s Brattle Group Prize in Corporate Finance in 2022, found that Black and Hispanic borrowers are less likely to gain from the use of machine learning to make decisions about creditworthiness. In the classroom, he seeks to ensure that AI aids intellectual inquiry rather than constraining it.

“I think there’s a real question for us of how to teach these tools and make sure students are still learning,” he said. “If you can get AI to do everything for you, do you have to do any thinking? How do we make sure not to fall into that trap?”

“All AI, All the Time”

Kyle Jensen has been exploring the potential of machine learning and AI for years. Before joining SOM, he co-founded Rho AI, a motorsports AI company acquired by GM. Since 2023, he has taught Large Language Models, SOM’s first course to focus exclusively on LLMs, with K. Sudhir, the James L. Frank ’32 Professor of Private Enterprise and Management. “As an organization that teaches management for business and society, it’s important to include AI everywhere,” he says.

But even he has been astounded by recent developments in the field, such as agentic AI. “The AI can reach out into the world and do things on your behalf,” he says. Since AI agents became available, “I’m just all AI, all the time, every day.”

Jensen uses the technology personally as well as in his teaching and research. He says he probably has between 15 and 20 agents operating on his laptop at any given time (usually pi, Codex, or Claude). But like others, he has concerns about its impact on society, noting that AI could result in “a pretty substantial change in humanity.”

Ma agrees, arguing that the current moment calls for subject-matter experts who can identify and grapple with the specific risks AI creates.

“It’s right to worry,” he says. “But we want to be productively worrying—identifying what we are worried about and developing a world and system to govern that, rather than saying, ‘This is scary; we’re not going to let people use it.’”

Ma plans to engage in that process, including in his new role as a Young Global Leader of the World Economic Forum, in which he will focus on advancing the responsible use of AI. The technology will have a profound effect on how all of us work, he adds.

“I think we’ll need to give up the word ‘worker,’” he says. “Everybody is going to be a creator; AI is going to be the worker. That’s the moment we’re in, and I’m glad to be witnessing it.”