SchrijfBlik: Safeguarding the Validity of Writing Assessment in the Age of AI

The rise of generative artificial intelligence (GenAI) in education necessitates new methods for assessing student writing. Teachers struggle to distinguish original work from AI-generated content, which challenges the validity of assessments. This paper introduces SchrijfBlik, a prototype designed to help secondary school teachers in the Netherlands gain insight into how students use GenAI for writing assignments. We follow the Double Diamond design process to explore and define our problem setting, and then develop a prototype to address the challenge of safeguarding validity in the age of AI. The results from a questionnaire among roughly 400 teachers and interviews with assessment experts highlighted the importance of visualizing process data on student-AI interactions in an intelligible way. SchrijfBlik was iteratively refined to create a user-friendly interface that shows prompts, edits, and reflections on a timeline, thus making the writing process of students more transparent for teachers. Early user tests suggest that SchrijfBlik can help to safeguard assessment validity, but further classroom evaluations are necessary to help us gain a better understanding of how we can best help teachers cope with the challenges of today’s AI world.