Conceptual Understanding and Conceptual Change Through Human AI Collaboration in Science Education: An Integrated Conceptual Framework from a Systematic Literature Review.

Human Artificial Intelligence (Human AI) Collaboration in science education is gaining momentum and more research is required on how it can facilitate students' conceptual understanding and conceptual change. Currently, most studies on AI in science education investigate the use of one particular tool and its effects on learning. However, it is the learning mechanisms that are triggered by these tools that help develop knowledge. This systematic literature review was conducted by performing a systematic search within the Scopus database which was last updated on 15 June 2026. In line with the PRISMA 2020 guidelines, a total of 28 empirical studies research were analyzed. These studies of empirical research were published between 2023 and 2026. Six roles of AI supporting the students' conceptual understanding as well as five mechanisms supporting the students' conceptual change were identified. Five interconnected learning mechanisms were derived from the analyzed studies, which are presented in an integrated conceptual framework. The framework is based on the theoretical framework of constructivism, the views on learning of the sociocultural theory, the approach of inquiry based science learning as well as on the metacognitive theory. Within the framework AI functions as a learning partner. The framework illustrates how, within science learning environments that are supported by AI, students continuously construct, evaluate and refine their scientific knowledge in order to achieve deeper understanding.