Impressions of ChatGPT: Using Survey Results to Inform AI Policy in Education

Abstract

ChatGPT is a powerful artificial intelligence chatbot developed by OpenAI. Because of the public availability and unprecedented writing capabilities of ChatGPT, it presents a unique policy obstacle for schools. On the one hand, ChatGPT and similar chatbots have the potential to be revolutionary learning tools, helping students to learn, study, and understand material. On the other hand, ChatGPT provides students with the means to cheat with unprecedented ease and competence, producing quality work in seconds with minimal understanding of the material required. In this paper, I examine the results of a survey I sent out to students, parents, and teachers in order to gauge policy opinions with regards to ChatGPT. The hope is that public opinion in this space can be used to inform policy making. I find that students are currently using ChatGPT at high rates, and that there is broad support for policy changes that can prevent AI cheating. Ultimately, I suggest that schools should minimize the use of take home assignments as a middle ground response to the growing risks of AI cheating.