Integrating chatbots in education: insights from the Chatbot-Human Interaction Satisfaction Model (CHISM)

Recent advances in Artificial Intelligence (AI) have paved the way for the integration of text-based and voice-enabled chatbots as adaptive virtual tutors in education. Despite the increasing use of AI-powered chatbots in language learning, there is a lack of studies exploring the attitudes and perceptions of teachers and students towards these intelligent tutors. This study aims to compare several linguistic and technological aspects of four App-Integrated Chatbots (AICs) and to examine the perceptions among English as a Foreign Language (EFL) teacher candidates. In this mixed-methods research based on convenience sampling, 237 college students from Spain (n = 155) and the Czech Republic (n = 82) interacted with four AICs over a month, and evaluated them following a rubric based on the Chatbot-Human Interaction Satisfaction Model. This scale was specifically designed to assess different linguistic and technological features of AICs such as response interval, semantic coherence, sentence length, and user interface. Quantitative and qualitative data were gathered through a pre-post-survey, based on the CHISM model and student assessment reports. Quantitative data were analyzed using SPSS statistics software, while qualitative data were examined using QDA Miner software, focusing on identifying recurring themes through frequency analysis. The findings indicated a moderate level of satisfaction with AICs, suggesting that enhancements in areas such as better adapting to learner needs, integrating interactive multimedia, and improving speech technologies are necessary for a more human-like user interaction.