Recent advancements in artificial intelligence (AI) have led to growing interest in understanding its potential applications and implications across various domains. Developed by OpenAI and released to the public in November 2022, ChatGPT has become widespread at an impressive speed, so much so that it reached one million users in five days. Reaching this number took 300 days for Facebook, 720 days for Twitter and 75 days for Instagram (Biswas, 2023; Fırat, 2023). ChatGPT’s wide range of use cases and its potential to improve the productivity of users in almost every industry are inspiring new conversations about this frontier AI application (Xames & Shefa, 2023). Education is among the most talked about. While some consider that this AI’s pioneering application will create a paradigm shift in various fields, including education (Bozkurt, 2023; Sallam, 2023), others emphasize the possible ethical challenges of ChatGPT and consider it a disruptive technology (Haque et al., 2023; Sardana et al., 2023). García-Peñalvo (2023) argues that the criticisms of ChatGPT stem from the resistance to change against its innovative and transformative potential rather than the disruptive nature of this technology. Since its public launch, ChatGPT’s ability to perform complex tasks in the field of education has caused mixed feelings among educators (Baidoo-Anu & Owusu Ansah, 2023).

The GPT-3 Natural Language Processing (NLP) model had 175 billion parameters, about ten times more than previous language models. GPT-3, an auto-regressive language model, was found to perform strongly on many NLP datasets, such as answering questions and completing missing words in the training process (Brown et al. 2020). While the reverberations of ChatGPT’s extraordinary success continued, its successor GPT-4 emerged and started to exhibit numerous new features. GPT-4 is more reliable, more creative, and can handle much more nuanced instructions than GPT-3.5 (OpenAI, 2023). The main differences between GPT-3.5 and GPT-4 are the parameter size of the models (GPT-3.5 has 175 billion parameters while GPT-4 has much more context length, the ability to use images as input in addition to text, and the use of Rule-Based Reward Models in its training (Koubaa, 2023; OpenAI, 2023).