Since going live in October 2022, ChatGPT, the artificial intelligence chatbot, has brought conversations about the ethics, utility, and accuracy of AI text generators to the fore. It has also started parallel conversations on the threat that AI poses to certain occupations.
In academic circles, these conversations seem to revolve around the incorporation of ChatGPT in the classroom and how this will affect student integrity. At present the opinions seem to fall on a continuum between embracing the technology and banning it completely. Some faculty want to encourage the use of ChatGPT as a way to keep up with rapidly changing technology (McMurtrie, 2023). Some faculty want to completely ban its use in any type of learning environment because it signals the end of critical thinking (Metha, 2023). These past few months have shown us that ChatGPT and its contemporaries are here to stay, and will probably be widely used by students.
So what have we as academic librarians learned about ChatGPT?
ChatGPT works on programmed knowledge. Therefore, originality, creativity and innovation are terms which cannot be applied to any of the works produced by the chatbot. This means that it may be very easy to detect a plagiarized ChatGPT assignment based on context and relevance (Marr, 2023). We have discovered that students are inadvertently learning a key information literacy skill: how to refine key terms by asking ChatGPT to produce an output tailored to their specific needs.
ChatGPT does not provide accurate information. The creators of ChatGPT have acknowledged this fact on their website when discussing the limitations of the chatbot (Kim, 2022). We have also learned from first hand experience that the chatbot may create its own sources and citations instead of citing existing credible sources. Even worse, it creates a mashup of invented citations and actual citations, thus giving the user a false sense of confidence.
Artificial Intelligence. (2022). Would a better reference for the abstract be if I were to cite you as the author? ChatGPT. https://chat.openai.com/auth/login.
In text citation may be as follows:
“The abstract is a hypothetical example that I generated based on my understanding of the topic” (Artificial Intelligence, 2022).
(Updated July 13, 2023). According to McAdoo (2023) of the APA Style Blog,
“The reference and in-text citations for ChatGPT are formatted as follows:
OpenAI. (2023). ChatGPT (Mar 14 version) [Large language model]. https://chat.openai.com/chat
- Parenthetical citation: (OpenAI, 2023)
- Narrative citation: OpenAI (2023)”
I, Rebecca Paulraj, do affirm that this blog post was not created by or edited by an artificial intelligence text generator.
Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence. 88 Fed. Reg. 16190 (March 16, 2023) (37 CFR Part 202).
Marr, B. (2023, Mar 3). The top 10 limitations of ChatGPT. Forbes (Online). https://www.forbes.com/sites/bernardmarr/2023/03/03/the-top-10-limitations-of-chatgpt/?sh=786f9f408f35
McAdoo, T. (2023, April 7). How to cite ChatGPT. APA Style. https://apastyle.apa.org/blog/how-to-cite-chatgpt
McKendrick, J. (2022, Dec 21). Who ultimately owns content generated by ChatGPT and other AI platforms? Forbes (Online). https://www.forbes.com/sites/joemckendrick/2022/12/21/who-ultimately-owns-content-generated-by-chatgpt-and-other-ai-platforms/?sh=5f2ee30f5423
McMurtrie, B. (2023, March 31). ChatGPT is already upending campus practices. Colleges are rushing to respond. Chronicle of Higher Education, 69(15), 11.
Mehta, R. (2023, May). A ban on ChatGPT does more harm than good. MIT Technology Review, 126, 20-21.