What Should Assessment Measure When AI Can Generate Answers?
What Should Assessment Measure When AI Can Answer Instantly?
Artificial intelligence can generate answers in seconds. This raises an urgent question for educators: what should assessment measure when producing an answer is no longer the hard part?
This Think Tank reimagines assessment as more than a tool for checking correct responses. It explores how carefully designed questions can make student reasoning visible — particularly in secondary ESL classrooms, where language difficulty and cognitive challenge must be balanced with care.
The session is grounded in classroom-based practice with Thai ESL learners aged 14–18. A conventional reading assessment showed what students remembered. A later assessment, using linguistically accessible but cognitively richer questions, began to reveal how they reasoned — asking students to notice evidence, connect ideas, challenge their first answers, justify interpretations, reflect on their thinking, and create original responses.
The core principle: the goal is not to make English harder — it is to make thinking more visible.
Participants will work with a practical six-stage framework — I AM → Notice → Connect → Challenge → Create → I AM — and examine how question architecture can support reasoning, metacognitive literacy, resilience, and learner confidence in an AI-saturated world.
This is not a session about using AI to produce faster answers. It is about what educators must now deliberately protect and cultivate: reasoning, reflection, communication, and the judgment to evaluate meaning — human capacities AI cannot supply on a student’s behalf.
Participants will leave with practical question stems, a set of visible-reasoning categories, and a classroom-ready method for redesigning assessments to reveal not just what students know, but how they think.
Facilitated By
Thembi Buthelezi
Secondary ESL Teacher / Classroom Researcher, Future Gen Leaders / Bangkok Secondary ESL Classroom Project
Thembi Buthelezi is a Secondary ESL Teacher and Classroom Researcher based in Bangkok, Thailand. Her work sits at the intersection of English language learning, human reasoning, metacognition, student identity, and AI-era education. Before moving into education, she built a 25-year international career across South Africa, the United States, China, Japan, and Thailand, including professional equities trading across NYSE, NASDAQ, and AMEX environments and advisory work with senior leaders in high-consequence business settings.
In her current classroom practice, Thembi works with Thai ESL learners aged 14–18 and is developing Making Student Thinking Visible, a classroom-based research project exploring how question architecture can reveal student reasoning beyond recall. Her work asks how educators can design linguistically accessible but cognitively rich learning experiences that strengthen human reasoning, metacognitive literacy, resilience, and learner confidence in the age of artificial intelligence.
