Block AI Tools on School Networks
Without Blocking Learning

# DNS Filtering Platform — Add AI Tools Blocklist as a Custom Integration
# Step 1: Navigate to Policies → Policy Components → Integration Lists
# Step 2: Click "Add New" → "Custom Feed URL"

# Feed file hosted on your own web server, regenerated daily
# from the API CSV (see the sync script further down)
Feed URL:  https://filter.district.internal/ai-domains.txt
Refresh:   3600 # seconds (hourly refresh)
Action:    Block
Category: AI Tools

# Step 3: Create or edit a DNS Policy
# Step 4: Under "Security Settings" → "Integration Lists"
#          → Enable the "AI Tools Blocklist" list
# Step 5: Apply the policy to the "Student-VLAN" identity

# Verification — test from a student device:
$ nslookup chatgpt.com
# Expected: NXDOMAIN or Umbrella block page

$ nslookup claude.ai
# Expected: NXDOMAIN or Umbrella block page

$ nslookup quillbot.com
# Expected: NXDOMAIN or Umbrella block page

# Content Filter Admin — Custom AI Tools Block List
# Navigate to: Admin Console → Filtering → Custom Lists

# Option A: Manual CSV Upload (for districts preferring manual review)
1. Download the CSV: ai_tools_database_sample.csv
2. Extract the "domain" column into a newline-separated text file
3. In your content filter admin → Custom Lists → "Create New List"
4. Name: "AI Tools Blocklist"
5. Action: "Block"
6. Paste or upload the domain list
7. Assign to: "Students" OU

# Option B: Automated sync via Google Workspace script
# This Apps Script fetches the latest feed and updates your content filter

function syncAIBlocklist() {
  const feedUrl = "https://www.aitoolsblocklist.com/api/database/?action=download_database";
  const response = UrlFetchApp.fetch(feedUrl, {
    headers: { "X-API-Key": "YOUR_API_KEY" }
  });
  // CSV columns: domain,category,subcategory — take the first column, skip header
  const domains = response.getContentText().split("\n")
    .slice(1).map(line => line.split(",")[0]).filter(Boolean);
  Logger.log("Fetched " + domains.length + " AI tool domains");

  // Push to your content filter via its API
  // See your filter vendor's API docs for custom list endpoints
}

# Example: Scheduled feed switching with a cron job on your DNS filter
# This script switches between testing and normal feeds

#!/bin/bash
# /opt/scripts/ai-blocklist-schedule.sh

API_URL="https://www.aitoolsblocklist.com/api/database/?action=download_database"
# Download today's CSV once (columns: domain,category,subcategory)
curl -s -H "X-API-Key: $API_KEY" "$API_URL" -o /tmp/ai_tools.csv

# Check if today is a testing day (dates stored in a config file)
TODAY=$(date +%Y-%m-%d)
if grep -q "$TODAY" /etc/ai-blocklist/testing-dates.txt; then
    # Testing period — block every AI-tool domain
    tail -n +2 /tmp/ai_tools.csv | cut -d, -f1 > /var/lib/dns-filter/ai-blocklist.txt
    echo "[$(date)] Testing period — full AI block active"
else
    # Normal period — block selected categories only (client-side filter)
    awk -F, 'NR>1 && ($2=="Text & Language" || $2=="Education & Learning" || $2=="Agents & Automation") {print $1}' \
        /tmp/ai_tools.csv > /var/lib/dns-filter/ai-blocklist.txt
    echo "[$(date)] Normal period — selective AI block active"
fi

# Reload the DNS filter (example for Pi-hole)
pihole restartdns reload-lists

echo "[$(date)] Feed updated: $(wc -l < /var/lib/dns-filter/ai-blocklist.txt) domains"