Columbia Business School Professor Says Academia’s Embrace of ChatGPT Can Do More Good Than Harm

Since the explosion of generative AI onto the scene last November, ChatGPT has been making waves across various sectors and industries. 

In the education space, the popular chatbot developed by OpenAI put instructors on high alert, with the New York City Department of Education blocking access to the technology in January of this year — the ban has since been lifted — over concerns of academic plagiarism and negative impacts on student learning. 

Dan Wang, associate professor of business at Columbia Business School, says that while there was little time for educators to adjust to the sudden rise of ChatGPT, encouraging his students to use ChatGPT to do their assignments has been more beneficial than harmful. 

“There was an opportunity, in many ways, to embrace this as a potentially beneficial use case in the classroom, while at the same time learning about its risks and limitations,” he told PYMNTS in an interview, adding that the only requirement is for students to disclose its use when submitting an assignment. 

And surprisingly, the students’ response, despite authorization to use the tool, was below his expectations. “I expected every student to be curious enough to want to embrace these tools but that didn’t actually materialize,” he remarked. “Throughout the semester, only about 40% of students ultimately adopted AI tools in their written assignments at least once.” 

The lesson from that one semester experience, he said, is that there needs to be a systemic change in how policymakers approach the use of AI tools in the classroom. “It has to be built in to justifiably be used as a tool for learning and one that’s creatively and also organically integrated into assignments and classroom exercises in ways that are obvious as well.” 

AI Requires More Student Effort

According to Wang, the use of AI tools like ChatGPT requires iteration and experimentation, exposing students to different perspectives that they would otherwise not have come up with on their own, while enhancing critical thinking and encouraging a more comprehensive understanding of the subject matter. “That’s something that’s not easy to do,” he said. “On average, students who use generative AI tools wind up spending more time on their homework.”

What that means is that educators must think about different standards in evaluating students, considering, for example, the effort when generative AI tools are involved, he added. 

In terms of preparing students for the future, Wang suggested that the education industry should align learning with the emerging skills and knowledge needed in the workforce, while arguing that there will be “greater value” placed on generalists who can synthesize diverse inputs generated by AI tools than functional specialists.

AI Impact on Learning Is Uncertain for Now

When it comes to the concerns among educators around the ethical issues ChatGPT pose to the integrity of academic and student learning, Wang said it would be premature to develop solutions that are systemic in nature without first understanding the real effects on the learning environment. 

“We can’t develop policy and then standardize and implement that policy without a good understanding of how students are using some of these tools,” he argued. 

This is especially important due to the dearth of data regarding how AI impacts learning, he added, and whether the use of tools such as ChatGPT exacerbate or bridge inequalities in student learning outcomes. 

As Wang said, “That’s something that we just have to establish first before coming up with any kind of solution to mitigate the potential harms of generative AI in learning environments.” 

Finally, for colleagues in academia who are still on the fence about the use of AI in the classroom, he said embracing these tools is not only necessary but also an opportunity to test its limits through their own lived experiences. 

“Unless we have evidence that guides us, a lot of our reasoning just rests on assumption. And so, in many ways, it’s an opportunity for instructors, educators, researchers to either confirm or disconfirm the skepticism that might put them on the fence about these types of technologies,” Wang said.