Generativ AIs roll i formativ bedömning : Användning av Generativ AI för individualiserad och formativ feedback i svenska och biologi på gymnasiet ur ett pedagogperspektiv.

Sammanfattning

The purpose of this study is to explore how generative artificial intelligence (GAI) can support individualized and formative assessment in Swedish upper secondary education. The study is grounded in a sociocultural perspective on learning and the feedback model by Hattie and Timperley. Data was collected through interviews with a project group of teachers and an open survey. Two teacher groups are analyzed: those teaching Swedish and biology, and those teaching other subjects. The results show that GAI has the potential to streamline the feedback process, particularly through quick and text-specific responses. Teachers of Swedish and biology used GAI directly in assessment situations, while other teachers mainly applied it as a tool for planning or inspiration. However, several limitations were identified: the GAI often lacked subject depth and struggled to connect feedback to learning goals or development strategies. Ethical concerns related to student over-reliance, academic integrity, and trust in GAI-generated feedback were also raised. The conclusion is that generative GAI can support assessment as a pedagogical tool, but only when interpreted through the lens of teacher expertise and ethical reflection. Further research is needed to explore how GAI can foster deep learning and meaningful formative assessment across subjects.