ChatGPT in classrooms is prompting educators to rethink assessment.

A qualitative study of 28 Canadian post-secondary educators finds generative AI such as ChatGPT is reshaping teaching and assessment, and participants outline five design principles — including explicit expectations and process-focused evaluation — to guide assessment in a post-plagiarism era.

Generative artificial intelligence is now present in higher education and is changing how students and educators work and assess learning. A recent qualitative study interviewed 28 educators across Canadian universities and colleges, from librarians to engineering professors. The authors also reviewed 15 years of research on AI and academic integrity, which shows both risks and benefits. Educators in the study are focusing on how to distinguish assistance that supports learning from assistance that substitutes for it. Key findings: - The study combined interviews with 28 Canadian post-secondary educators and a review of research spanning 15 years. - AI tools can produce human-like text and may reproduce errors or social biases, which can make misconduct harder to detect. - AI can also provide legitimate support, for example for students with disabilities or those learning an additional language. - Participants identified three assessment boundary areas: prompting, critical thinking and writing. - The group proposed five design principles: explicit expectations; process over product; tasks that require human judgment; developing evaluative judgment; and preserving student voice. Summary: The research frames a post-plagiarism context in which co-creating with AI challenges traditional assumptions about authorship while keeping honesty and integrity important. Participants and authors suggest institutions can update policies and training and design assessments that foreground human judgment and the learning process as next steps.