Leveraging Generative AI

01 Jun 2024

Leveraging Generative AI

HBS navigates the use of these transformative tools for teaching and learning

Re: Mitchell B. Weiss (Richard L. Menschel Professor of Management Practice Chair, MBA Required Curriculum); Jeffrey J. Bussgang (Senior Lecturer of Business Administration); Chiara Farronato (Glenn and Mary Jane Creamer Associate Professor of Business Administration); By: Jennifer Gillespie

Topics: Technology-Artificial Intelligence Innovation-Technological Innovation Education-Business Education Education-Curriculum and Courses Information-Data and Data Sets

Mitchell Weiss (Photo by Susan Young)

Four decades after HBS became the first business school in the country to require the use of personal computers in the MBA Program, the School is undergoing a different kind of technological transformation, one that leverages generative artificial intelligence (GenAI) for both teaching and learning.

Last year, the MBA Program introduced ChatGPT as a component of the required course Data Science for Managers and the elective Launching Tech Ventures (see below). Faculty members have integrated it into other courses as well, and the new elective Generative AI for Business Leaders both educates students about the business implications of this new technology and encourages hands-on experimentation with it (see story).

“We want students to become able, inventive, and responsible leaders in their application of these tools, and we’re inviting their use of them while they’re at HBS,” explains Mitchell Weiss, the Richard L. Menschel Professor of Management Practice and chair of the Required Curriculum in the MBA Program.

Read more about

From Chalkboards
to Chatbots

A large language model such as ChatGPT is a type of GenAI that is trained on extensive text data and capable of understanding and generating human-like text. HBS is equipping students with the skills and information needed to use AI-powered platforms such as ChatGPT for case preparation—to clarify complex concepts or analyze data—and to reflect more deeply on classroom discussions. A special module offered to students, “Learn with Baker Library: Generative AI for MBAs,” guides them through the effective and responsible use of GenAI tools to support their learning. In addition, the MBA Program is exploring the viability of AI-assisted tutoring.

Weiss has seen firsthand the effects of employing these tools, recalling one student’s comment that “This is going to change the way I view my entire career.” Moments of insight like this are happening now at HBS and, Weiss observes, “There’s a role for the School and its faculty to play in helping members of the HBS community see how transformative these tools can be.”

While acknowledging that HBS is actively exploring the use of GenAI for teaching and learning, Weiss says, “I never want us to come across as naïve to the ethical, moral, and existential concerns about generative AI. What we’re doing is encouraging students to learn about these tools and decide what they should and shouldn’t be used for.”

In making GenAI available to students, both the University and the School have established guidelines to preserve the rigor and integrity of the educational experience while fostering creative and ethical use of these tools to enhance learning. Students may upload a case to authorized, non-public platforms to further their understanding of the subject matter; can use GenAI tools in some classes, when appropriate; and can draw upon them to prepare for exams.

In addition to helping students navigate the use of GenAI to enhance their learning, HBS faculty are pursuing research into the opportunities and challenges facing businesses in today’s AI world, and are leveraging GenAI for case development and classroom exercises. Weiss has experimented extensively with ChatGPT-4 to outline the structure of a case, anticipate questions students might ask, and even fine-tune a syllabus. While he doesn’t suggest turning over case writing to chatbots, Weiss says they can be valuable for gaining broader perspectives on case dilemmas and preparing for class discussions.

He also uses ChatGPT-4 in the classroom. For Weiss’s Public Entrepreneurship course, he conducted an AI exercise called “Storrowed” with students, using GenAI tools, to understand why some trucks ignore the height warnings of overpasses on Boston’s Storrow Drive and regularly get stuck, causing traffic backups. The goal was to find out why and where else this was happening, what kind of economic impact these mishaps have, and what solutions could be found to prevent them in the future. “Students were asking the chatbot ‘Can you do a five-whys analysis, can you do a fault tree analysis’—tools and frameworks from other classes—to help solve the problem. It was fascinating to see,” says Weiss. One group of students proposed legislation that would shift liability to the rental truck companies, and with the chatbot’s help, devised the acronym SHIFT for the Safety Height Information for Trucking Act.

Looking more broadly beyond the use of GenAI for a classroom exercise or to prepare for an exam, Weiss says: “I hope ultimately that we will educate students who can make a positive difference with AI in the world—that people will look to them for leadership on this topic, and to HBS as a place for how you can use these tools to elevate teaching and learning."

For the “Storrowed” exercise, students drew on ChatGPT to understand why trucks keep getting stuck on Storrow Drive.

GENERATIVE AI IN THE CLASSROOM

Included below are two examples of the ways in which GenAI is being used in the MBA Program to enhance the educational experience for students.

DATA SCIENCE FOR MANAGERS
HBS introduced Data Science for Managers (DSM) into the Required Curriculum in 2023 to provide students with the insights needed to lead data-driven organizations. Students were given access to ChatGPT Plus to assist them with data analysis and coding assignments, which also enabled faculty members to see the impact of such tools on learning.

To get a better sense of how often students in her sections of DSM accessed the tool and what they thought about it, Chiara Farronato, the Glenn and Mary Jane Creamer Associate Professor of Business Administration, did her own analysis. She found that about half of the class never or rarely used it, saying that they wanted to figure out the concepts themselves. The rest of the class used it sometimes or frequently, primarily to understand statistical or machine learning concepts or to get ChatGPT’s help with coding. “Using these AI models was almost as if I had a private tutor in the room with me,” one student remarked, while others said the tool was not as reliable as they had hoped, causing them to lose trust in its capabilities. Nevertheless, Farronato sees it as an opportunity for further exploration: “GenAI will dramatically speed up the way in which we can teach managers how to build and lead a data-driven organization. A large barrier to making that happen so far has been the need for our students to gain basic familiarity with at least one coding language. GenAI is removing that barrier: We can now start ‘coding in English,’ leaving the translation effort to GenAI. All managers are left to do is envision the insights their data could provide and strategize how to leverage them for business success.”

LAUNCHING TECH VENTURES
For his second year entrepreneurship elective, Launching Tech Ventures (LTV), Senior Lecturer Jeffrey Bussgang conducted an innovative experiment last fall through which he created an AI faculty copilot to help him teach. Bussgang and two former students developed the AI chatbot they called ChatLTV, which was trained on a comprehensive body of course-related materials—roughly 15 million words—and historical Q&As from the course’s Slack channel. To ensure the security of the content and respect copyright laws, the chatbot was developed in a walled-off environment using OpenAI’s ChatGPT-4 and Microsoft’s Azure OpenAI Service.

ChatLTV enabled the roughly 250 students in the three sections of the course to engage with the platform either privately, with only the student and Bussgang able to see the interactions, or publicly, so all in the course could see. Bussgang found the chatbot to be a useful tool, providing him with a unique view into what students were curious about before class, thereby enhancing the quality of in-class discussions. Students also found it to be an invaluable resource for course preparation, using it to ask a wide range of questions about case studies, analysis, and administrative matters. “I could use it to check my answers but more importantly understand if my methodology was directionally correct,” said one student. “It was almost like having a professor by my side.”

—JG

For HBS and Harvard AI-related resources, please visit alumni.hbs.edu/GenAI-Resources.