Abstract
The recent launch of ChatGPT by OpenAI has created a profound global impact, initiating deep questions among educators about how it might affect education, syllabi and teaching methods. Currently, the full scope of potential benefits and risks associated with ChatGPT in education remains unclear, given that its impact surpasses the level of preparation educators and institutions may have had for such a pre-trained generative AI tool. While Artificial Intelligence in Education has long been a subject of research, with a particular focus on developing Intelligent Tutoring Systems, the emergence of ChatGPT marks a distinctive advancement in this field. Unlike dedicated Intelligent Tutoring Systems, ChatGPT is readily available to a diverse spectrum of educational stakeholders, including teachers, students, schools, universities, and educational institutions. Scholars have initiated assessments of ChatGPT's effectiveness across various educational disciplines, even though ChatGPT was not explicitly designed for educational purposes. However, the widespread accessibility of ChatGPT, coupled with its extensive knowledge base, necessitates the development of comprehensive evaluation frameworks. In this paper, we introduce a holistic evaluation framework tailored for ChatGPT. This framework takes into account both soft and hard skills, and it is designed to seamlessly incorporate ChatGPT into Intelligent Tutoring Systems, making it suitable for a wide range of educational fields. By establishing a connection between ITS and ChatGPT, as they are both AI tools, we can benefit from the substantial background work achieved by previous research in ITSs to evaluate the educational influence of ChatGPT.