Generative AI Conception of the Nature of Science

10 Pages Posted: 20 Dec 2023 Last revised: 15 Apr 2024

See all articles by Matthew Nyaaba

Matthew Nyaaba

University of Georgia - College of Education

Date Written: December 12, 2023

Abstract

Humans encounter numerous challenges in comprehending the nature of science, attributable to factors such as educational limitations, cognitive biases, the inherent complexity of science, and the rapid pace of advancements. This study assessed the extent to which generative AI (GAI) tools such as ChatGPT and Gemini understand and articulate scientific concepts, considering their current usage by teachers and students. Structured as a descriptive survey, the Form-C of Lederman et al.'s (2002) instrument, which is specifically designed to measure views on the nature of science among teachers and pre-service teachers was used to prompt the GAI tools. The responses were then summarized and analyzed based on Lederman's dichotomy, which classifies views as either naive or informed. The findings revealed that both GPT-4 and Gemini possess an informed understanding of the nature of science, recognizing its complexity, the variability of scientific knowledge, and the evolutionary nature of scientific theories and methodologies. This outcome contrasts with existing literature, which implies that human subjects, including teachers and pre-service teachers, often demonstrate a blend of naïve and informed views regarding the nature of science. The informed views shown by the GAI tools suggest a level of insight that potentially surpasses the average understanding among teachers and pre-service teachers. Moreover, despite both ChatGPT and Gemini presenting informed views, their responses, and the examples they provided varied in scope and details. Gemini offered more comprehensive answers to certain prompts, whereas GPT-4 was a bit more concise. Additionally, the way each tool supported its responses with examples also varied, though both were adept at appropriately addressing the prompts posed. This study, therefore, highlights the need for further research into the implications and the development of GAI tools to support science education.

Keywords: Gemini, VNOS, GPT-4, Informed View, Teachers, Pre-service teachers

JEL Classification: A20, A22, A23, I20, I26, I23, O33, O31, O30, Q55

Suggested Citation

Nyaaba, Matthew, Generative AI Conception of the Nature of Science (December 12, 2023). Available at SSRN: https://ssrn.com/abstract=4661602 or http://dx.doi.org/10.2139/ssrn.4661602

Matthew Nyaaba (Contact Author)

University of Georgia - College of Education ( email )

United States

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
321
Abstract Views
1,290
Rank
184,908
PlumX Metrics