Rethink student assessment as AI enters everyday classroom practice
Across the Eastern Caribbean, teacher educators are calling for a reset in how students are assessed. The message: move from memorisation to application, innovation, and ethical use of technology.
At the Eastern Caribbean Joint Board of Teacher Education's two-day meeting at the Sagicor Cave Hill School of Business and Management, the theme was clear-Repositioning Teacher Education as Engaged Scholarship. AI is already here; assessment must catch up.
Why assessment must change
Director of Tertiary Education, Dr Roderick Rudder, told educators that traditional testing no longer fits a classroom where students use AI for research and assignments. "We cannot avoid artificial intelligence. It's an important tool to aid in learning, to aid in teaching," he said.
His push: judge how learners apply knowledge, solve real problems, and create useful outputs. Students are already using AI for School-Based Assessments (SBAs); our assessments need to be more authentic with higher order questioning.
What to assess now
- Application and transfer: Can students use knowledge to solve a new problem in a new context?
- Ethical AI use: Are they transparent about prompts, tools, and limitations? Do they cite AI support?
- Process over product: Can they show their steps, revisions, and decision points?
- Critical thinking: Can they question outputs, detect bias, and verify claims?
- Collaboration and feedback: Can they give and use critique to improve work?
- Local relevance: Can they design solutions that fit their school, home, or community environments?
Practical assessment ideas you can deploy this term
- AI-augmented projects: Require a brief "AI usage log" with prompts used, tools consulted, and what was kept, edited, or discarded-plus why.
- Oral defenses: Short viva-style interviews where students explain choices, test assumptions, and respond to new constraints.
- Live performance tasks: Timed "build, analyze, or explain" activities where students can use approved tools and must justify outputs.
- Versioned portfolios: Drafts, critiques, and final work with reflection on how feedback and AI shaped the outcome.
- Case analyses: Compare AI-generated answers with expert sources; identify errors, bias, and missing context.
- Community challenges: Tackle a real issue at school or in the neighborhood and present a tested solution.
- Structured peer review: Rubrics for accuracy, clarity, ethics, and local fit; students revise after peer feedback.
Sample rubric criteria for AI-enabled tasks
- Clarity of problem definition and constraints
- Quality of reasoning and evidence
- Transparency of AI use (tools, prompts, limits)
- Original contribution beyond AI output
- Local relevance and feasibility
- Accuracy, citation, and verification
- Reflection on what worked, what failed, and next steps
Higher order question stems you can reuse
- How would you test whether this solution holds up under a tighter budget or time limit?
- Where could this approach fail in our community, and how would you redesign it?
- Which source or method gives you the strongest evidence, and why?
- What did the AI miss, and how did your thinking fill the gap?
- If you had to improve this with half the data, what would you change first?
Classroom norms for responsible AI
- Declare AI use: Students attach a short statement listing tools and prompts used.
- Cite and verify: Cross-check AI claims with reliable sources; flag uncertainties.
- Protect privacy: No personal or sensitive data in prompts.
- Value process: Grades reward reasoning, iteration, and reflection-not just final polish.
- Bias check: Ask students to test outputs against different groups and contexts.
Teacher preparation must evolve
Dr Rudder urged teacher training institutions to modernise curricula with a stronger blend of theory, practice, and effective use of AI. The goal: more graduates who are work-ready and able to contribute to social and economic development.
- Integrate AI literacy into methods courses and practicum.
- Adopt micro-assessments that test authentic skills, not recall.
- Run lesson study cycles featuring AI tools, then share what works.
- Co-design school projects with industry and community partners.
- Support students who are "smart" in non-traditional ways-highlight problem-solving seen outside class (gaming, fixing devices, community projects) and convert it to creditable learning.
30-60-90 day plan for schools
- Next 30 days: Publish an AI-use policy, set citation norms, and pilot oral defenses in one unit.
- Next 60 days: Convert one exam or SBA component into a performance task with an AI usage log and reflection.
- Next 90 days: Build a shared bank of higher order questions and rubrics; run a showcase of student projects solving local problems.
Useful resources
The bottom line
Assessment has to move from recall to real-world use, ethical tool use, and community impact. Students are already working with AI; our job is to set clear guardrails and reward thinking, not copying.
As Dr Rudder put it, the task is to see how learners "apply the knowledge that is generated to solve problems and to come up with innovative solutions" in their schools, homes, and communities. Start small, make the process visible, and grade what matters.
Your membership also unlocks:
