The rapid rise of AI-generated text in student submissions has resulted in a technological arms race: students using AI in submissions, institutions adopting policing tools and AI detection, and students responding by using so-called ‘humanizer’ tools to circumvent detection. The result is a systemic erosion of trust for written assignments and submissions. What’s clear is that to verify and validate student writing skills, new approaches are necessary. This session details a successful, large-scale pilot at an Australian university to reimagine writing assessment. In the course, Writing for Science - BMS100, faculty assess scientific writing as an outcome. To protect integrity and verify student writing, the faculty shifted their focus from the final product to the writing journey itself using Moodle as the assessment platform. The team implemented a layered validation framework that included: in-class assessment periods, engaging students about the goals, modeling expectations for student conduct during the class time, making Moodle the assessment platform where writing occurred and leveraging the Cursive plugin for TinyMCE to capture revisions and identify/manage pasted
- Teacher: Thibault Joseph