Safieh Moghaddam, Associate Professor, Language Studies, Centre for Teaching and Learning
Dina Soliman, Educational Developer, Digital Pedagogies
Filtering the Noise: Tools, Trends, and Tensions
Generative AI can introduce “noise” into assessment: polished text that masks learning, unclear authorship, overconfident claims supported by weak or fabricated evidence, and tasks that inadvertently reward fluency over intended learning outcomes. This interactive workshop supports instructors in redesigning one existing assignment so the “signal” (reasoning, evidence use, and decision-making) becomes visible and assessable, whether or not students use GenAI.
The session opens with a brief “noise map” activity: participants identify where GenAI most interferes with assignment effectiveness (e.g., product-over-process, unclear contribution, unverifiable claims, misalignment with outcomes, equity/hidden advantages). Results are surfaced quickly (e.g., via polls) to identify shared pain points.
Next, the facilitator demonstrates a brief before/after assignment makeover. The “before” version highlights where AI noise can creep in (broad designs, vague expectations for evidence, no process visibility). The “after” version shows how three design moves reduce noise and amplify the signal: visible thinking (students show reasoning and choices), verification/evidence (students substantiate and check claims), and a brief process note (students document decisions and tool use, if any).
Participants then complete an Assignment Makeover Lab, using a guided template to revise their own assignment (or a provided sample). In peer consult pairs or small groups, they refine drafts using a short checklist and add one transparent AI-use pathway (AI as a brainstorming partner, critic/editor, or comparator). Participants leave with a concrete, implementable draft assignment design.
Leave A Comment
You must be logged in to post a comment.