University of Texas Agriculture professor Jared Mumm recently put in motion a plan to prevent students from using the artificial intelligence chatbot, known as ChatGPT, for their term papers. Mumm used the chatbot as a “rat” – he copied and pasted the essays he’d received into the chatbox, and ChatGPT replied that more than half of the writings were “not his”. The professor was quick to suspend the supposedly fraudulent students, without any evidence. The accused students were wrongfully deprived of their diplomas, so the incident had to be thoroughly investigated.
According to the person familiar with the case, ChatGPT is a sophisticated tool and its use must be approached carefully. It is true that it is capable of detecting AI-composed material, but it cannot identify whether a certain document was written by a real person or not. The Reddit user Delicious_Village112 conducted an experiment to prove this: He asked ChatGPT whether Professor Mumm’s PhD dissertation was generated by an AI and the chatbot confirmed that “The text contains several characteristics that are consistent with the content generated by AI”.
The University of Texas Agriculture is now taking the necessary steps in order to investigate the whole fiasco. They are developing policies that should prevent similar cases in the future and the students will meet the professor individually to discuss the situation. This incident clearly showed that ChatGPT can be dangerous in the hands of a careless teacher, so it is now even more important to know how to use AI-based technology in education.