How to include ChatGPT within business school teaching

By Yoram (Jerry) Wind, Lauder Professor Emeritus and Professor of Marketing, The Wharton School at University of Pennsylvania

In January, the New York City education department blocked access to ChatGPT on the city’s school devices and networks. Some educators and school administrators fear the app, which uses artificial intelligence to convert prompts into high-quality text, might lead to widespread cheating, as students pass off ChatGPT’s work as their own. Haunted by similar concerns, the school districts of Los Angeles, Oakland, Seattle, and Baltimore have instituted similar bans, as have districts in Alabama and Virginia. Other school districts are pondering similar restrictions.

Are we right to worry about the “negative impacts on student learning,” as the NYC education department calls them?

Based on more than five decades of teaching experience at the Wharton School of the University of Pennsylvania, I consider the recent decision by the New York City Board of Education and other school districts to be a mistake. While it is possible that text generators such as ChatGPT or Microsoft Bing AI could be used to plagiarise or cheat when we assess students based on testing for memorisation, why are we addicted to archaic testing approaches and trying to protect them? What does anyone gain by glorifying learning by rote?

How to include ChatGPT within education

I believe that a much better approach is to assess the students' understanding of the material they are required to learn. During my time teaching at Wharton, I have never given a closed book in-class exam that relied on memory. All my exams have been open book or take-home exams that allowed students to use any source in demonstrating their understanding of the material.

If we prevent students from using tools that can make them faster and more efficient at expressing what they have learned, it is as silly as arguing that students will become better writers if they are barred from using typewriters and computers or that using calculators will hamper their efforts to master math.

In the case of ChatGPT and other generative AI platforms that follow or will follow them, the solution is simple: We should encourage students to start with the ChatGPT response to the query they are trying to study. Then they should be evaluated based on their ability to critically analyse the algorithm’s response, creatively improve it, and ideally implement what they have learned.

To be sure, this means that teachers will have to work harder. This new approach to assessment will require them to read and evaluate each response. It will also eliminate the convenience of automatic grading of multiple choice questions. However, at the same time, it will encourage students to focus on understanding rather than memorising their lessons. It will also increase the relevance of and their engagement with the topics they are studying.

This solution is consistent with history’s lessons. You cannot stop technological advancement by mandating it to disappear. We cannot decree that students can no longer use calculators, typewriters, computers, and mobile phones. Such bans never work -- and the blockage of large language models like ChatGPT will ultimately fail. It is for similar reasons that prohibition did not work, and most anti-drug legislation is ineffective.

Those in charge of our education system need to realise that they should not fight technological advances. Instead, they need to embrace these new technologies. It is much better to adapt pedagogical systems so that they can work with innovative technologies and not against them.

Teachers, for the sake of our students, it is time to get real.

Photo by Matheus Bertelli