Building Fair AI Policies

Why Blanket Bans Don't Work

Many schools initially responded to ChatGPT with outright bans. These have largely failed because:

Enforcement is Impossible:

• Students use personal devices
• AI text can be modified to avoid detection
• Detection tools are unreliable
• Policing creates adversarial relationships

Bans Miss the Point:

• AI is here to stay in education and work
• Students need to learn responsible use
• Complete prohibition prevents teaching opportunities
• "Just don't use it" isn't preparation for the real world

Bans Create Inequity:

• Students with means access AI anyway
• Those following rules are disadvantaged
• Underground use continues without guidance
• Honest students are penalized

The Tiered AI Policy Framework

Instead of binary "allowed/banned," use a tiered approach that matches AI use to learning goals:

LEVEL 1: AI RESTRICTED
AI use not permitted for this assignment
Focus: Skill building, authentic assessment

LEVEL 2: AI GUIDED
AI permitted with disclosure and reflection
Focus: Learning with AI, transparency

LEVEL 3: AI OPEN
AI use encouraged as a learning tool
Focus: AI literacy, critical evaluation

Each assignment specifies its level, creating clarity for students.

Level 1: AI Restricted

When to Use:

• Skills that must be developed independently
• Baseline assessments of individual capability
• High-stakes evaluations
• Early skill-building before AI assistance

Examples:

• In-class timed writing
• Math problem-solving tests
• First drafts of analytical essays
• Foreign language composition
• Foundational coding exercises

How to Communicate:

"This assignment is AI Restricted (Level 1). You may not use ChatGPT, Claude, or similar AI tools. This is because [learning goal]. You'll have opportunities to use AI on other assignments."

Enforcement:

• Supervised conditions when possible
• Process-based assessment
• Oral follow-up to verify understanding
• Clear consequences for violation

Level 2: AI Guided

When to Use:

• Students learning to work with AI appropriately
• Assignments where AI can support but not replace thinking
• Building transparency habits
• Teaching AI literacy

Examples:

• Research paper with AI assistance documented
• Code with AI-generated sections identified
• Writing revision with AI feedback acknowledged
• Problem sets where AI use is disclosed

Requirements:

1. AI use must be disclosed
2. Original thinking must be evident
3. AI contributions must be identified
4. Reflection on AI use required

Disclosure Template:

"In this assignment, I used [AI tool] for [specific purpose]. The AI helped me with [what it did]. I then [what you did with AI output]. My original contributions include [your thinking]."

How to Communicate:

"This assignment is AI Guided (Level 2). You MAY use AI tools with these requirements: [list requirements]. This approach helps you [learning goal]."

Level 3: AI Open

When to Use:

• Teaching AI literacy itself
• Real-world simulation where AI is standard
• Creative exploration
• Efficiency is prioritized over process

Examples:

• "Use AI to draft, then critically revise"
• "Generate three options with AI, then choose and improve"
• "Fact-check AI output using primary sources"
• "Collaborate with AI on creative project"

Still Requires:

• Critical evaluation of AI output
• Documentation of AI interaction
• Original judgment and decisions
• Understanding of AI limitations

How to Communicate:

"This assignment is AI Open (Level 3). You are encouraged to use AI tools. However, you must [requirements]. The goal is [learning objective about AI use]."

Writing Your Classroom Policy

Step 1: Establish Philosophy State your core beliefs about AI in education:

"In this class, I believe AI is a tool that, used thoughtfully, can enhance learning. My goal is to prepare you to use AI responsibly, not to pretend it doesn't exist."

Step 2: Explain the Framework Describe your tiered approach:

"Different assignments will have different AI policies. Each assignment will clearly state whether it's AI Restricted (no AI), AI Guided (AI with disclosure), or AI Open (AI encouraged). This isn't arbitrary—each level supports specific learning goals."

Step 3: Define Consequences Be clear and fair:

"Using AI when not permitted, or failing to disclose AI use when required, is an academic integrity violation. Consequences include [specific consequences]. However, I am more interested in learning conversations than punishment. If you're unsure, ask first."

Step 4: Provide Examples Help students understand:

"An essay first draft (Level 1) helps me understand your current thinking—AI would hide that. A research project (Level 2) lets you use AI for efficiency, but I need to see your judgment. An AI literacy exercise (Level 3) teaches you to work with AI critically."

Sample Classroom Policy Document

AI POLICY FOR [CLASS NAME]

Philosophy: AI tools like ChatGPT are part of our world.
Learning to use them well is a skill I want you to develop.
But some learning requires your unassisted effort.

The Framework:
Each assignment specifies its AI level:
- 🔴 Level 1 - AI Restricted: No AI use permitted
- 🟡 Level 2 - AI Guided: AI permitted with disclosure
- 🟢 Level 3 - AI Open: AI use encouraged

Requirements:
Level 1: Complete work independently. AI use = integrity violation.
Level 2: Disclose AI use using the disclosure template.
         Show your original thinking.
Level 3: Use AI freely but document and evaluate its contributions.

Disclosure Template (Level 2):
"I used [tool] for [purpose]. AI helped with [what].
My original contributions: [your thinking]."

Consequences:
First violation: Conference + redo assignment
Repeated violations: Grade reduction + parent contact
Egregious violations: Referred to administration

Questions?
If you're unsure whether AI use is appropriate, ASK.
I'd rather you ask than guess wrong.

Communicating with Parents/Guardians

Parents have questions about AI. Address them proactively:

Sample Parent Communication:

"Dear Families,

You may have heard about AI tools like ChatGPT and wondered how they affect your student's education. Here's my approach:

I don't pretend AI doesn't exist. Instead, I'm teaching students to use it responsibly—knowing when it helps, when it hinders learning, and how to use it ethically.

My policy uses three levels: some assignments restrict AI (to build skills), some guide AI use (with disclosure), and some encourage AI use (to develop AI literacy).

You can support this by:

• Asking your student what AI level their assignment has
• Encouraging them to disclose AI use when required
• Discussing AI's role in their future careers

Questions? Please reach out.

[Teacher Name]"

Department/School Alignment

Individual policies work better with coordination:

Benefits of Alignment:

• Students have consistent expectations
• Teachers share resources and strategies
• Policies are harder for students to game
• Professional learning community develops

How to Align:

1. Share your policy with colleagues
2. Discuss common framework adoption
3. Create shared disclosure templates
4. Coordinate on tool recommendations
5. Regular policy review meetings

Common Policy Mistakes

Mistake 1: Policy Without Explanation Students need to understand why, not just what.

Mistake 2: Inflexible Levels Some assignments may need flexible approaches. Build in room for teacher judgment.

Mistake 3: All Punishment, No Conversation Learning from mistakes matters more than punishing them.

Mistake 4: Ignoring AI's Benefits A policy focused only on preventing misuse misses teaching opportunities.

Mistake 5: Set and Forget AI evolves rapidly. Review policies each semester.

Policy Review Checklist

Use this to evaluate your policy:

Element	Question	Check
Clarity	Can students understand what's allowed?	☐
Fairness	Are consequences proportionate?	☐
Flexibility	Does it allow for different assignments?	☐
Enforceability	Can you actually enforce it?	☐
Educational	Does it teach, not just police?	☐
Transparent	Do students know why it exists?	☐
Updated	Is it current with AI developments?	☐

Key Takeaways

1. Blanket bans fail because they're unenforceable and miss teaching opportunities
2. The Tiered Framework (Restricted, Guided, Open) matches AI use to learning goals
3. Each assignment specifies its level—clarity prevents confusion
4. Disclosure requirements teach transparency without banning AI
5. Consequences should be fair and focused on learning, not just punishment
6. Communicate with parents proactively about your approach
7. Review policies regularly as AI capabilities change

:::