AI Literacy in the NTU Core Curriculum

CC0007 sits within NTU's interdisciplinary Core Curriculum and asks students to examine how science and technology shape human life. Public reporting on NTU's AI initiative identifies Science and Technology for Humanity as the mandatory course in which responsible AI and AI-agent skills are taught.

The institutional setting emphasises inquiry-driven experiential learning, project work, collaborative knowledge building and ethical, data-informed use of AI.

Within that setting, my contribution has been to co-create and deliver the AI-literacy curriculum at tutorial level: translating broad pedagogical ambitions into weekly activities, research workflows, assessment guidance, feedback and classroom facilitation.

Generative AI makes fluent answers abundant. That changes the educational problem. If students are assessed mainly on producing an answer to a predefined question, AI can complete much of the visible task while leaving the most important learning invisible.

The curriculum therefore has to move from solving predefined problems toward framing complex inquiries.

Students need to learn how to decide what is worth investigating, assemble trustworthy evidence, reconcile perspectives, identify the limits of an AI-generated proposal and take responsibility for a course of action.

The learning design joins two elements that are often separated:

The intended result is not faster content production. It is stronger critical awareness, interdisciplinary collaboration and a more explicit human–AI division of cognitive work.

Agency is scaffolded rather than assumed. Students progress through three levels, with support reducing as the inquiry becomes more open-ended.

1. 01

   Asynchronous preparation

   Foundational readiness

   Students establish a baseline in source-grounded research, prompt design and responsible tool use through curated guidance before the live session.

2. 02

   Synchronous inquiry

   Facilitated application

   Weekly cases turn the classroom into a place for critique and iteration. Students compare outputs, expose weak assumptions and improve how they question, steer and verify AI.

3. 03

   Project-based synthesis

   Applied mastery

   Teams use grounded research and agentic workflows to address complex human-centred challenges, while documenting the decisions that must remain with people.

The assessment sequence makes the inquiry process visible across collective and individual work.

1. 01

   Research and frame

   Teams build a multi-perspective source library, critically evaluate AI-assisted research and define a defensible problem statement.

2. 02

   Develop and communicate

   Students create a grounded interdisciplinary response and translate it into a visual argument and a stakeholder-focused oral pitch.

3. 03

   Synthesize and reflect

   Each student explains how their understanding changed, what the group learned across disciplines and where human intervention altered the AI-supported process.

This progression separates three capabilities that a polished final answer can otherwise conceal: framing a problem, developing a justified response and reflecting on how judgement changed.

Different AI systems are assigned different cognitive roles rather than offered as a generic toolbox.

The pairing matters: agents can expand the search and reasoning process, while a bounded evidence base helps keep conclusions inspectable and verifiable.

• Data provenance: curating credible sources and tracing claims back to evidence.
• Output verification: moving from plausible generation to evidence-based academic work.
• Agentic inquiry: progressing from passive search and retrieval to active questioning, critique and iteration.
• Human intervention: recording where human logic overrode, redirected or constrained an AI-supported process.
• Collaborative knowledge building: asking, creating, connecting and improving ideas across disciplinary perspectives.
• Responsible judgement: deciding what to delegate, what to verify and what must remain a human decision.

Across project work, the difficult parts are rarely producing a plausible solution. The recurring challenges are more fundamental:

• shallow problem framing and premature convergence on an answer;
• weak evaluation of source quality and AI-supported claims;
• limited integration across disciplinary and stakeholder perspectives;
• insufficient attention to trade-offs, feasibility and implementation context;
• low visibility of the student's own intervention and judgement.

The curriculum responds by making the process of inquiry observable and by assessing decisions that fluent output can otherwise conceal.

Across four consecutive semesters, I led 42 CC0007 tutorial classes reaching approximately 1,700+ unique undergraduate students. Student Feedback on Teaching across these classes averaged approximately 4.3/5, with recurring strengths around feedback quality, approachability, enthusiasm, participation and analytical thinking.

Scale makes coherence important. The three-tier structure provides a repeatable rhythm—prepare, practise, apply—while leaving room for students to pursue different problems, sources and stakeholder perspectives.

• Structure and agency must develop together. Open-ended inquiry works better when students first have a reliable process for grounding, critique and iteration.
• Verification needs an artifact. Source libraries and intervention records turn “critical thinking” from an aspiration into visible practice.
• AI literacy is relational. Students learn from comparing approaches, challenging each other's assumptions and building stronger explanations together.
• Assessment directs attention. If framing, evidence and reflection are not rewarded, students rationally focus on the polish of the final product.
• Human judgement must be named. Students need explicit practice deciding what to delegate, what to verify and what they remain responsible for.

The practical shift is from treating AI literacy as prompt technique to treating it as a curriculum of judgement. Assessment must reward the quality of the question, the integrity of the evidence trail, the rationale for human intervention and the student's ability to defend a decision—not merely the fluency of the final artefact.

AI literacy is the capability to reason, learn and act effectively when AI becomes part of the cognitive system.

Last reviewed September 2026.