AMBA's Ambition magazine: Issue 61, March 2023

OPINION 

stimulate students to reflect on the knowledge they have acquired to build more detailed and complete answers, such as debates, case studies and collaborative group projects. The last of these in particular opens avenues for professors to guide students’ use of AI tools, helping them understand the technology’s limitations and principles for how to use it most effectively to support – but crucially, not replace – their thinking. Teaching staff at business schools and other higher education institutions should maintain a balance between the use of new AI technologies such as ChatGPT and traditional practices, including lectures, practical exercises, projects, readings and presentations. They should also provide regular feedback and assessments for students to ensure they have a well-rounded understanding of the topics discussed in classes. Then there is the question of exams. Put simply, if an AI is capable of passing the exam, the exam should be changed. Assessments should increasingly focus on testing skills and competencies and steer away from being a trial of how much knowledge each student can remember. Going forward, we must focus more on teaching students to understand, rather than simply recall, information. With this approach in place, and clearly communicated punishments for those caught cheating using AI technologies, there is little reason for schools and universities to ban tools like ChatGPT from teaching practices. We must remember how

sky-rocketed in the wake of ChatGPT’s launch, a number of posts and articles began to appear, expressing worries that generative AIs could be used to cheat on exams or solve homework problems. It is true that the mishandling of ChatGPT and similar technologies could result in impoverishing students’ thinking. While the AI is a valuable tool in gathering and processing information, it cannot replace the teaching of a professor, nor the critical self-reflection of a student. This is because it does not truly think like a human. It operates in an ingenious but regurgitative fashion, speed-harvesting knowledge from other sources, for which it often does not include proper citations. Tasks must be set for the AI in clear terms, allowing it to use the wording of the question as a jumping-off point. This is what we call ‘the art of the prompt’. If an instruction is too imprecise or poorly phrased, ChatGPT’s answer will be mechanically ‘stupid’. One example of a clear command is this: “Explain quantum physics to me as Einstein would have done to a 15-year-old.” It is, therefore, important to ensure that students don’t just copy the answers generated by ChatGPT, but rather use them as a starting point for thinking about more detailed and complete answers. The technology is a means, not an end. This type of approach can be encouraged by promoting exercises and activities that

“It is important to ensure students use answers generated by ChatGPT as a starting point to think about

transformative the introduction of the first internet search engines was in

the 1990s. Generative AIs have

the potential to be similarly revolutionary for numerous sectors of the