Do generative artificial intelligence (GenAI) and science education mix? A systematic review of the literature
Kason Ka Ching Cheung, Amina Zerouali, Jenna Koenen, Sibel Erduran
Abstract
Generative Artificial Intelligence (GenAI) has been incorporated in different contexts of science education, such as K-12 education and teacher education. However, science educators have raised concerns about the dissonance between what GenAI can offer and what science education aspires to teach to learners. For example, GenAI may undermine social learning goals, learning of science, and epistemic understanding of science. Using a systematic review approach, we examined the variables, data collection tools, and features of teaching interventions in 22 empirical studies to trace how GenAI is integrated into science education. Critically, we identified the extent to which research studies reflect synergy as well as dissonance between GenAI and science education. In so doing, we explored examples of fruitful integrations of GenAI in science education. Our findings showed that variables, data collection tools, and teaching interventions focused on students’ understanding of either science or GenAI. We propose a framework that guides the unification of diverse learning goals in relation to disciplinary practices in science education in the context of GenAI. Based on this framework, we propose future research and teaching interventions that harness synergy as well as dissonance to foster students’ and teachers’ higher-order thinking while using GenAI in learning science or developing their pedagogical competence, respectively.