An Interview with ChatGPT on Emergency Remote Teaching: A Comparative Analysis Based on Human–AI Collaboration

Abstract

Background/purpose -ChatGPT, a recent form of AI-based language model, have garnered interest among people from diverse backgrounds with its immersive capabilities. Using ChatGPT to support or generate scientific research has also created an ongoing debate over its advantages versus risks. The present study aimed to conduct an AI-enabled research process using ChatGPT so as to evaluate its potential to generate an accurate, clear, concise, and unbiased information as these are essential elements of rigorous scientific work. Materials/methods -To achieve this aim, we worked on emergency remote teaching (ERT), which garnered significant interest due to its wide-spread use during the COVID-19 pandemic, and created opposing views, particularly in comparison to online teaching. We conducted a simultaneous query on ChatGPT-3.5 and 4 on five basic themes: (1) the definition and emergence of ERT, (2) the appropriateness of ERT for different grade levels, (3) a comparison between ERT and online teaching, (4) the possible outcomes of ERT, (5) the future prospects and uses of ERT, and we performed a comparative evaluation of these responses with regard to accuracy, clarity, conciseness, and potential bias. We also used Cohen's kappa to assess inter-rater agreement in our analysis. Results -The results indicated that both versions were capable of generating accurate information without significant bias although the responses lacked depth and insight with being somewhat repetitive. As the level of judgment required by the query increased, the performance of ChatGPT-4 was much better; it provided clearer and more concise responses with a more synthesized and detailed categories of information on ERT. Conclusion -Based on our results, we state that the cooperation of human and artificial intelligence is still warranted to ensure an accurate and reliable output from AI-based scientific queries. If ChatGPT is a plane with innovative technologies, there still needs to be a pilot in the cockpit to make use of these technologies in the best way so as to fly the plane safely to its destination.

Using Generative Artificial Intelligence Tools to Explain and Enhance Experiential Learning for Authentic Assessment

“…In this way, this work considers GenAI tools not merely as objects, but as study subjects, because of their capacity to provide human-like responses and seamlessly engage in conversations with humans [23]. For instance, previous works in artificial intelligence (AI) show the possibility of interviewing ChatGPT for information elicitation or obtaining insights into diverse research topics [42][43][44][45].…”

ChatGPT and its ethical implications for STEM research and higher education: a media discourse analysis

“…Noticeably, as of May and through September 2023, OpenAI has launched ChatGPT-4, which alarms academics to rethink the rapid evolution of AI chatbots and its language models, despite its claims about creation, safety, and benefits for all of humanity (Barash et al, 2023;Lewandowski et al, 2023;Peters et al, 2023b;Tülübaş et al, 2023; see also OpenAI, 2015OpenAI, -2023. Despite the existing body of literature on AI chatbots in general and the recently growing body of evidence regarding the advent of ChatGPT, far too little empirical research has been conducted in this area.…”

“…More recently, Tülübaş et al (2023) tested the Human-AI collaboration to generate scientific research projects via ChatGPT-3.5 and 4. The investigators interviewed ChatGPT on emergency remote teaching (ERT) and generated texts regarding the research themes and evaluations.…”

A Conversation with ChatGPT about Digital Leadership and Technology Integration: Comparative Analysis Based on Human–AI Collaboration

“…The developments also helped reduce its likelihood of generating hallucinated or incorrect information, mentioned as a serious limitation of ChatGPT-3.5 (OpenAI 2023b). Recent research conducted after the release of ChatGPT-4 (e.g., Teebagy et al 2023;West 2023;Tülübaş et al 2023) also indicates that ChatGPT-4 demonstrates capabilities far beyond its previous version, even in showing emotions (Zhao et al 2023), performing inductive reasoning and inferring peoples' feelings or perspectives (Michail et al 2023), despite acknowledging that there is still much room for improvement if it is expected to act like a human.…”