An integrated model for predicting pupils' acceptance of artificially intelligent robots as teachers

Artificially intelligent robots as teachers (AI teachers) have attracted extensive attention due to their potential to relieve the challenge of global teacher shortage and realize universal elementary education by 2030. Despite mass production of service robots and discussions about their educational applications, the study of full-fledged AI teachers and children's attitudes towards them is quite preliminary. Here, we report a new AI teacher and an integrated model to assess how pupils accept and use it. Participants included students from Chinese elementary schools via convenience sampling. Questionnaires (n = 665), descriptive statistics and structural equation modeling based on software SPSS Statistics 23.0 and Amos 26.0 were carried out in data collection and analysis. This study first developed an AI teacher by coding a lesson design, course contents and Power Point with script language. Based on the popular Technology Acceptance Model and Task-Technology Fit Theory, this study identified key determinants of the acceptance, including robot use anxiety (RUA), perceived usefulness (PU), perceived ease of use (PEOU) and robot instructional task difficulty (RITD). Moreover, this study found that pupils' attitudes towards the AI teacher, which could be predicted by PU, PEOU and RITD, were generally positive. It is also found that the relationship between RITD and acceptance was mediated by RUA, PEOU and PU. This study holds significance for stakeholders to develop independent AI teachers for students.