ChatGPT Threatens Educational Progress

Photo courtesy of DeepMind

Released in November 2022 by OpenAI, ChatGPT is an artificial intelligence (AI) chatbot that can perform various natural language processing tasks such as answering academic questions, producing creative works, generating computer programs and more. Its conversational capabilities led to its prominence across various academic institutions such as DePauw University, providing an efficient solution to common student issues. However, the long term learning process of students could be affected by the misuse of this popular AI chatbot, as it threatens to replace foundational university experiences.

According to a study conducted by Radboud University and published in the Journal of Product Innovation Management, ChatGPT utilizes transformer-based models that can rapidly generate ideas across different fields of study. For instance, knowledge extraction is accelerated by the chatbot as it can summarize long bodies of text, providing users with generalized information on any topic from world history and fairy tales to modern day romance. By producing paragraphs that seem to be written by humans, students may be inclined to claim ChatGPT’s work as their own. This is reflected by a survey conducted by last January, which found that one in three college students used ChatGPT to accomplish their written academic works, with 60 percent of respondents relying on the chatbot for more than half of their school responsibilities.

This is highly concerning for a liberal arts university such as DePauw, due to its curriculum’s emphasis on hands-on projects such as research, case studies, essays, and creative works for different academic disciplines. Since AI writing has the tendency to develop biased perspectives from a pre-existing database, the innovation process of students for their academic pursuits can be hindered by the generative ideas from ChatGPT.

Despite the powerful capabilities of the AI chatbot, its calibration and training process did not come without its faults. According to an investigative report by Time Magazine, OpenAI exploited a team of Kenyan workers who were paid less than $2 daily to curate the content of the chatbot to eliminate topics such as domestic violence, sexual abuse, and other criminal acts. This was essential since ChatGPT’s predecessor, GPT-3, was more prone to vulgar vocabulary due to its exposure to a wider training dataset of internet-based words that were not properly cleaned by data scientists. The controversial development and flawed composition abilities of ChatGPT taint the credibility of this widely-used AI tool.

However, the construction of new texts from ChatGPT’s natural language processing model can still make it difficult for professors to distinguish between the academic output of a student from that of an AI chatbot. Based on a recent statement from Noam Chomsky, a professor from the University of Arizona who is considered a founder of modern linguistics, ChatGPT is “basically high-tech plagiarism” that heightens an individual’s aversion to learning. Since it has the ability to construct student-like writings with barely any grammatical errors, the ideas presented by the chatbot can be harder to identify through common plagiarism detector tools. This can raise significant concerns within DePauw’s Code of Conduct, especially when students’ unethical practices are left undetected.

Students must exercise caution when using influential AI tools such as ChatGPT for their academic pursuits, as it can impact their educational experience inside and outside the classroom. We must not let a chatbot destroy the rich culture of critical thinking, creativity, and academic exploration that has upheld the foundation of the DePauw curriculum for more than 18 decades and will serve as the heart of the student community in the years to come.