Skip to main content
    Tech Horizon Academy Logo
    AI Governance

    AI Frameworks &
    Systems

    Practical governance frameworks to help your team use AI safely and effectively

    Data Classification

    Traffic Light Protocol

    A simple system to classify what data can be used with AI tools

    🔴 RED LIGHT
    Strictly Prohibited

    Data that triggers Privacy Act:

    • Tax File Numbers (TFNs)
    • Biometric Data (facial recognition, voiceprints)
    • Sensitive Health Information
    • Unreleased IP/Trade secrets

    NEVER enter into cloud-based AI, even paid

    🟡 AMBER LIGHT
    Caution - Enterprise Only

    Internal confidential data:

    • Client Names & Addresses (only in Enterprise licenses)
    • Financial Reports (P&L, balance sheets)
    • Strategy Documents (marketing plans, board minutes)

    Safe in Enterprise tools, NOT in free tools

    🟢 GREEN LIGHT
    Go - Safe to Use

    Public/low-risk data:

    • Website Content (blogs, landing pages)
    • Generic Templates (email drafts, policy skeletons)
    • Public Market Data (competitor pricing, industry reports)

    Safe in any tool

    The 10% Rule

    Human-in-the-Loop Checklist

    AI does 90% of the heavy lifting, human does the final 10%. Use this checklist before sending any AI-generated content.

    Pre-Send Verification
    0 of 4 checks completed
    0%

    Verified specific dates, prices, names? (AI can hallucinate)

    Does this sound like us? (Remove "delve", "tapestry", "comprehensive landscape")

    Did I click the links? (AI sometimes invents URLs)

    Is language inclusive and appropriate for Australian context?

    Team Handout Template

    Plain English AI Policy

    A simple, readable policy template you can customize and share with your team

    AI Policy Template
    Ready to customize for your organization
    1

    Approved Tools Only

    List paid, secure tools only

    2

    The 'No-Training' Rule

    Never put work data into free tools

    3

    The Red Light Rule

    Never enter client names, TFNs, passwords

    4

    You Are The Pilot

    You are responsible for every word sent

    5

    Declare It

    If AI created substantial part, declare it

    Full Template Preview
    This is what gets copied to your clipboard
    PLAIN ENGLISH AI POLICY - TEAM HANDOUT

1. APPROVED TOOLS ONLY
   Only use paid, enterprise-grade AI tools approved by management.
   Current approved tools: [INSERT YOUR APPROVED TOOLS HERE]

2. THE 'NO-TRAINING' RULE
   Never put work data into free AI tools. Free tools may use your 
   data to train their models. Only use enterprise/paid versions.

3. THE RED LIGHT RULE
   NEVER enter into any AI tool:
   • Client names or addresses
   • Tax File Numbers (TFNs)
   • Passwords or access credentials
   • Biometric data
   • Sensitive health information
   • Unreleased IP or trade secrets

4. YOU ARE THE PILOT
   AI is your co-pilot, not autopilot. You are personally responsible 
   for every word that leaves this organisation. Always review, verify, 
   and approve AI-generated content before sending.

5. DECLARE IT
   If AI created a substantial part of any work product, declare it 
   to your manager/team. Transparency builds trust.

---
Questions? Contact: [INSERT CONTACT EMAIL]
Last updated: 29/09/2026