LTIC — AI Tool Use, At a Glance

A printable policy reference for when it's OK to put information into AI tools (Claude, Claude Code, ChatGPT, Copilot, etc.) when working at LTIC. Page 1 is the at-a-glance summary (traffic-light rules + helpful patterns); page 2 is vendor-specific facts (Claude, Gemini, ChatGPT) with source links.

Traffic-light zones and vendor cards reference UBC's Information Security Standards: ISS U1 (data classification — Low / Medium / High / Very High) and ISS U9 (contractual controls for third-party services with Medium-risk data). Keep these references current if Prism updates the framework.

Edit

Everything lives in ltic-ai-guidelines.html — single self-contained file, no build step. Open it in a browser to preview locally.

Common places to edit:

• The three categories — <ul class="examples"> blocks (one per Red / Yellow / Green row)
• The hint list — <ol class="tips"> block
• Claude Code factual card — <section class="cc-card"> near the bottom
• Stamp (v0.1 · draft, owner, review cadence) — top-right of the header

Refresh the preview image

After substantive edits, regenerate ltic-ai-guidelines.png:

"/Applications/Google Chrome.app/Contents/MacOS/Google Chrome" \
  --headless=new --disable-gpu --hide-scrollbars \
  --force-device-scale-factor=1.5 --window-size=900,1180 \
  --screenshot="$PWD/ltic-ai-guidelines.png" \
  "file://$PWD/ltic-ai-guidelines.html"