Generative AI & Academic Integrity: IP and JC Guardrails for 2026

TL;DR
Generative AI is welcome in Singapore classrooms only when students can cite the tool, verify outputs, and reflect on learning. Anchor your routines on MOE's 2024 advisory, adhere to SEAB's 2025 examination rules, and document every AI touchpoint in coursework logbooks and IB/AL submission forms.

1 Policy Guardrails You Must Internalise

1.1 MOE's 2024 Generative AI in Education advisory

• Purpose: Encourage meaningful, ethical use of AI tools while safeguarding data protection and academic honesty. (MOE guidance)
• Key expectations:
  • Always disclose AI assistance in logbooks, reflection journals, or submission coversheets.
  • Use AI as a brainstorming or drafting partner, not a final-answer machine.
  • Apply school's data-protection policies; avoid uploading personal identifiers or exam questions into public AI systems.
• Teacher authority: Schools can restrict AI usage for specific tasks. Ignoring instructions counts as misconduct.

1.2 SEAB's 2025 examination rules

• SEAB's 2025 "Rules and Regulations for School Candidates" explicitly bans unauthorised digital assistance during national exams. (SEAB rules PDF)
• Any AI-generated content submitted without acknowledgement can be treated as plagiarism or cheating.
• Pre-exam declarations require candidates to certify that coursework (SPAs, H1/H2 projects) reflects authentic effort; AI misuse risks disqualification.

1.3 IB's Academic Integrity Policy (2022 update)

• The IB policy defines generative AI output as a form of "third-party assistance" that must be cited like any other source. (IBO Academic Integrity)
• For EE/IA submissions, students must document AI use in the RPPF (Reflections on Planning and Progress Form) or methodology sections.
• Supervisors may request raw prompts or conversation logs to verify authenticity.

2 Cite–Verify–Reflect Workflow (Printable Routine)

2.1 Step-by-step workflow

1. Cite
   • Record the tool name, model version, date accessed, and prompt in your research log.
   • Example: ChatGPT (OpenAI GPT-4o mini), accessed 19 Sep 2025 — Prompt: “Outline energy audit experiment variables for H2 Physics IA.”
2. Verify
   • Cross-check AI-generated facts against primary sources (MOE syllabi, SEAB data booklets, peer-reviewed articles).
   • Highlight validated vs unresolved claims using colour codes.
3. Reflect
   • Write 2–3 bullet insights about what the AI output contributed (idea generation, rephrasing, error flagging) and what you still need to do manually.
   • Log emotional/mental load to monitor over-reliance.

2.2 Classroom-ready tracker

| Task                    | AI tool         | Prompt summary                  | Manual checks                               | Reflection                                                   |
| ----------------------- | --------------- | ------------------------------- | ------------------------------------------- | ------------------------------------------------------------ |
| H2 Physics SPA planning | ChatGPT 4o mini | Suggested apparatus list        | Verified with SEAB SPA guide (p.12)         | AI helped brainstorm; still need teacher vetting             |
| GP essay outline        | Claude 3 Sonnet | Generate argument counterpoints | Cross-checked with Straits Times commentary | Useful for idea expansion; rephrased final paragraphs myself |

📎 Tip: Store this log in a shared Google Sheet with your subject tutor so they can monitor compliance quickly.

3 Scenario Playbook for IP & JC Classrooms

3.1 Mathematics WA preparation

• Allowed: Using AI to propose alternative solution paths, then manually verifying steps against textbook methods.
• Not allowed: Submitting AI-generated solutions without personal working.
• Teacher check: Request raw working in pen-and-paper or digital stylus to prove conceptual ownership.

3.2 Science investigations (IB IA / H2 Practical)

• Use AI to generate hypothesis variants or apparatus comparisons, but you must validate against SEAB practical guides.
• Keep screenshots of prompts and highlight subsequent lab results to illustrate divergence.

3.3 Humanities essays

• AI may assist with brainstorming angles or structuring outlines.
• Final drafting requires personal synthesis with citations from credible sources (e.g. MOE policy briefings, parliamentary debates).
• Use plagiarism detection tools (Turnitin) to ensure paraphrasing is authentic.

3.4 Project-based assessments

• Create an AI usage appendix in Google Docs or Notion that attaches prompts, outputs, and annotations.
• Encourage group members to sign off on each AI interaction so accountability is shared.

4 Monitoring & Governance Toolkit

4.1 Personal governance checklist

• Maintain an AI interaction log for each subject.
• Update a source triangulation map (AI claim → official document → action taken).
• Schedule weekly integrity reflection: what went well, what still feels blurry.

4.2 Teacher/parent oversight tools

• SLS analytics: Teachers can see when AI-assisted quiz attempts spike; follow up to ensure comprehension.
• Google Docs revision history: Spot sudden jumps in prose quality; request commentary to confirm authorship.
• AI detection heuristics: Instead of relying on unreliable detectors, examine voice consistency and require oral defenses.

4.3 Logging template (copy-ready)

- Week of: 22 Sep 2025
- Subjects monitored: H2 Chemistry IA, IP Math WA2 prep, EE draft 1
- Key AI aids used: ChatGPT 4o mini, Gemini 1.5 Pro
- Verification sources: SEAB practical rules (p. 14-18), MOE policy brief (2024), Nature Education article on kinetics
- Outcome: All AI-derived data re-run in personal spreadsheet; flagged one hallucinated reference and replaced with MOE bulletin

5 Well-being & Executive Function Guardrails

5.1 Managing cognitive load

• NIE research on socio-cognitive training shows that structured self-reflection improves executive function and reduces burnout. Encourage students to log AI usage alongside mood/energy indicators to surface overload trends. (NIE socio-cognitive study)

5.2 Healthy screen-time boundaries

• The American Optometric Association recommends the "20-20-20" rule (every 20 minutes, focus on something 20 feet away for 20 seconds) to curb digital eye strain. (AOA computer vision guidance)
• Build in device-free problem-solving blocks to ensure conceptual mastery without AI crutches.

5.3 Emotional resilience checkpoints

• Use HPB or school counsellor resources to discuss academic stress triggered by AI comparison culture.
• Adopt journalling prompts: "What did AI do for me? What did I learn myself? How confident am I without the tool?"

6 Deliverables & CTA

6.1 Printable "Cite–Verify–Reflect" worksheet

1. Task & deadline
2. AI tool name + version
3. Prompt(s) used
4. Verification source(s) + status (✓ / follow-up)
5. Reflection (skills gained, gaps to close)
6. Teacher/parent countersignature

6.2 Academic integrity escalation flow

1. Spot potential breach (sudden jump in writing voice, missing citations).
2. Collect evidence (AI logs, revision history, interviews).
3. Discuss restorative actions (redo sections, oral defense, academic honesty lesson).
4. Document resolution for future audits.

🎯 CTA: Download the editable AI usage tracker and register for our "Academic Integrity in the Age of AI" webinar. We guide IP/JC families through practical routines, policy updates, and wellbeing strategies that keep students future-ready without compromising ethics.

6.3 Reference stack

• MOE — Guidance on Generative AI
• SEAB — Rules & Regulations for School Candidates (2025)
• IBO — Academic Integrity Policy (2022)
• NIE — Socio-cognitive training research
• HealthHub — Mental wellbeing essentials