Launch Live Tool ·
Standards · Verification · Integrity
A comprehensive framework for ensuring AI-generated curriculum content is safe, accurate, and age-appropriate before it reaches students.
Launch Live Tool · Quick Start · Overview · Web Tool · Full Framework · Contributing
AI tools like ChatGPT, Gemini, and Claude are increasingly used by teachers and curriculum developers to generate lesson plans, assessments, and study materials. But AI-generated content carries unique risks in classrooms:
- Hallucinations — AI can produce confident but completely false statements that students accept as fact
- Age-inappropriate content — Output may contain themes, vocabulary, or complexity unsuitable for the target age group
- Curriculum misalignment — Generated content may not match national curriculum standards
- Cultural insensitivity — Content may reflect Western-centric perspectives inappropriate for Nigerian/African classrooms
- Data privacy violations — Student data entered into AI tools may be processed in ways that violate data protection laws
ACIF solves this by defining a structured verification pipeline — with risk classification, age-band checks, hallucination detection, and curriculum alignment gates — that every piece of AI-generated content must pass before reaching students.
| Component | Description |
|---|---|
| 4 Risk Tiers | Classify content from Low (Tier 1) to Critical (Tier 4) based on subject sensitivity, age group, and use case |
| 6 Age Bands | Developmental guidelines from Early Childhood (3–5) through Pre-Tertiary (18+) |
| Five-Gate Pipeline | Sequential verification gates: Pre-Delivery Screening → Factual Verification → Age-Appropriateness → Curriculum Alignment → Final Approval |
| FACT Method | Hallucination detection protocol: Find claims → Assess risk → Cross-reference sources → Tag results |
| Governance Structure | Roles, responsibilities, escalation paths, and audit requirements |
ACIF is aligned with four international and national standards:
| Standard | How ACIF Aligns |
|---|---|
| NERDC National Curriculum | Gate 4 enforces alignment with Nigeria's 9-Year Basic Education Curriculum and Senior Secondary curriculum |
| Nigeria Data Protection Act 2023 (NDPA) | Section 7 implements consent requirements, data minimisation, and prohibited data inputs for student protection |
| ISO/IEC 42001:2023 | Framework maps to AIMS clauses: risk assessment (6.1), operational planning (8.1), monitoring (9.1), and corrective action (10.2) |
| UNESCO GenAI in Education (2023) | Principles 1–8 implement UNESCO's guidance on human agency, age restrictions, inclusion, and AI competency development |
acif-repo/
├── README.md # This file
├── LICENSE # CC BY-SA 4.0 License
├── CHANGELOG.md # Version history
├── CONTRIBUTING.md # Contribution guidelines
├── CODE_OF_CONDUCT.md # Community standards
├── SECURITY.md # Security policy
├── docs/
│ ├── ACIF_Framework_v1.0.pdf # Full framework (41-page PDF)
│ ├── ACIF_Framework_Full_Text.md # Framework in Markdown format
│ ├── IMPLEMENTATION_GUIDE.md # Step-by-step adoption guide
│ ├── FAQ.md # Frequently Asked Questions
│ └── GLOSSARY.md # Key terms and definitions
├── web-tool/
│ ├── index.html # ACIF Verification Pipeline Tool
│ ├── styles.css # Styles
│ └── app.js # Application logic
├── templates/
│ ├── hallucination-log.csv # Blank hallucination log template
│ ├── content-registry.csv # Blank content registry template
│ ├── ai-tool-assessment.md # AI tool evaluation checklist
│ ├── gate-checklist.md # Printable Five-Gate checklist
│ └── parental-communication.md # Template letter for parents
├── assets/
│ └── acif-banner.svg # Project banner
└── .github/
└── ISSUE_TEMPLATE/
├── bug_report.md # Bug report template
├── feature_request.md # Feature request template
└── framework_feedback.md # Framework content feedback
- Read the Framework — Start with the Full Framework PDF or the Markdown version
- Follow the Implementation Guide — See docs/IMPLEMENTATION_GUIDE.md for a 3-month rollout plan
- Use the Web Tool — Use the Live Web Tool or open
web-tool/index.htmllocally - Print the Templates — Use the files in
templates/for offline verification workflows
- Clone this repository:
git clone https://github.com/Chukwuemerie-ezieke/acif-framework.git
- The web tool is a static HTML/CSS/JS application — no build step required
- Open
web-tool/index.htmlto run locally - Integrate the risk classification logic and gate checklists into your platform
- See CONTRIBUTING.md for how to contribute
The ACIF Verification Pipeline Tool is a browser-based application that implements the full framework:
| Module | Function |
|---|---|
| Dashboard | Overview stats, pipeline visualisation, quick actions |
| Risk Classifier | Interactive questionnaire to assign Risk Tier 1–4 |
| Verification Pipeline | Sequential Five-Gate checklist with reviewer sign-off |
| Age Checker | Age-band specifications with allowed/prohibited theme lists |
| Hallucination Log | Log, search, and export hallucination incidents |
| Content Registry | Track all approved AI-generated content with full audit trail |
Requirements: Any modern web browser (Chrome, Firefox, Edge, Safari). No server needed — all data is stored in browser localStorage.
To run: Simply open web-tool/index.html in your browser.
| Tier | Level | Examples | Required Gates |
|---|---|---|---|
| Tier 1 | Low | Timetables, creativity prompts, templates | Gates 1 + 5 |
| Tier 2 | Medium | Lesson plans, explanatory notes, objective assessments | All 5 Gates |
| Tier 3 | High | Health/science facts, history, religion, Early Childhood content | All 5 Gates + Senior sign-off |
| Tier 4 | Critical | Sensitive topics, formal exams, external publications | All 5 Gates + Director + Legal review |
| Band | Ages | Key Restrictions |
|---|---|---|
| Band 1 | 3–5 (Nursery) | Oral/visual only, max 8-word sentences, no conflict themes |
| Band 2 | 6–8 (Primary 1–3) | Lexile 200–500, no violence or fear-inducing content |
| Band 3 | 9–11 (Primary 4–6) | Lexile 500–800, death/illness handled with care |
| Band 4 | 12–14 (JSS 1–3) | Lexile 800–1100, puberty content requires Health Educator sign-off |
| Band 5 | 15–17 (SS 1–3) | Lexile 1100–1400, complex moral/ethical topics allowed |
| Band 6 | 18+ (Pre-Tertiary) | University-level, all WAEC/NECO content must match official syllabus |
A systematic approach to detecting AI hallucinations in educational content:
F — Find the claim Identify every factual assertion (not opinions or questions)
A — Assess the risk Is this claim specific enough to be wrong? Is it checkable?
C — Cross-reference Verify against at least 2 independent authoritative sources
T — Tag the result Mark as VERIFIED, MODIFIED (corrected), or REMOVED
No content proceeds to the next gate until all claims are tagged and zero unverified claims remain.
We welcome contributions from educators, developers, policymakers, and anyone working at the intersection of AI and education. See CONTRIBUTING.md for:
- How to propose framework changes
- How to contribute to the web tool
- How to submit translations
- Code of Conduct
This work is licensed under the Creative Commons Attribution-ShareAlike 4.0 International License (CC BY-SA 4.0).
You are free to:
- Share — copy and redistribute the material in any medium or format
- Adapt — remix, transform, and build upon the material for any purpose, including commercially
Under the following terms:
- Attribution — You must give appropriate credit to Harmony Digital Consults Ltd
- ShareAlike — If you remix or build upon this material, you must distribute your contributions under the same license
Developed by Harmony Digital Consults Ltd Anambra State, Nigeria
ACIF is part of Harmony Digital Consults' mission to make AI adoption in Nigerian education safe, ethical, and standards-compliant. The framework is designed for adoption by any educational institution in Nigeria and across Africa.
Contact: For institutional adoption support, training, or integration into EdTech platforms, reach out via the repository Issues page or contact Harmony Digital Consults directly.
Launch Live Tool · ACIF v1.0 · © 2026 Harmony Digital Consults Ltd · Licensed under CC BY-SA 4.0