Speaker
Description
Generative AI has intensified long-standing concerns about authorship, academic integrity and the validity of classroom-based writing assessment. Yet reliance on AI-detection tools may produce unreliable judgements while doing little to help learners understand how AI can be used critically and responsibly. This practice-based workshop proposes an AI-aware assessment cycle designed to protect evidence of independent language ability while developing learner agency and AI literacy in IELTS writing classrooms.
Grounded in principles of assessment validity, transparency and learner accountability, the proposed cycle treats AI use not as a binary choice between prohibition and unrestricted access, but as a traceable pedagogical process. It comprises six stages: controlled independent writing; explicitly bounded AI-assisted revision; prompt and revision logging; learner justification of accepted and rejected suggestions; oral verification; and rubric-based comparison of independent and revised performance. By triangulating the written product, the revision process and the learner’s explanation of their decisions, the framework seeks to distinguish independent ability from AI-supported performance without reducing academic integrity to surveillance.
During the workshop, participants will examine contrasting writing samples, identify potential threats to validity and authorship, and adapt the cycle to an assessment task from their own teaching context. They will also consider how familiar IELTS criteria—Task Response, Coherence and Cohesion, Lexical Resource, and Grammatical Range and Accuracy—can be used to document learning across different stages of AI-supported revision.
The workshop argues that meaningful AI literacy requires more than competent tool use. It requires learners to evaluate AI output, reject unsuitable suggestions, disclose assistance transparently and remain accountable for the final text. Participants will leave with an adaptable assessment protocol and practical templates for balancing validity, learner development and responsible human–AI collaboration.
