AI in STEM education: The relationship between teacher perceptions and ChatGPT use

The rapid development of AI in the recent years has triggered numerous discussions in the education sector since it offers numerous opportunities, but also challenges. This study addresses one context in which AI might become a relevant tool for different purposes: German secondary education. This study therefore examines the extent to which STEM teachers in German secondary education assess the benefits but also the risks of ChatGPT and how this influences the future use of AI in an educational context. As part of a questionnaire study, 102 STEM teachers were investigated. The current and future use of ChatGPT and various teacher variables were recorded with questionnaires based on classifications of researchers as well as political institutions. Relationships between the variables were explored using a path model investigating hypotheses regarding the affect heuristic, perceived risks and benefits as well as effects on teaching quality. In general, AI is still used relatively rarely, but future usage expectations are high compared to current usage. Results further revealed that perceived competence and benefits of ChatGPT have a positive influence on the use and intention to use it. Perceived risks and concerns have no significant influence on the usefulness of ChatGPT in the classroom, indicating that STEM teachers use AI in the classroom despite potential concerns and perceived risks. Additionally, perceived benefits and risks are negatively associated with each other, indicating that teachers rely on an affect heuristic when judging the usefulness of AI technology in the classroom.