Your first protected use in ChatGPT
Audience: users. Editorial review: 2026-09-15.
Use a made-up example that identifies nobody and contains no work data while getting familiar with Defend.
- Sign in to Defend and open ChatGPT.
- With the ChatGPT tab selected, open the popup and check protection.
- Prepare your fictional example and read any notices before confirming a send.
- If Defend asks you to review detections, check the result and follow the visible options. If a block or limited state appears, stop and resolve the notice.
Expected result
Section titled “Expected result”You understand the state and review steps. Do not use real information to test detection. A ChatGPT response does not prove that every sensitive item was detected.
More help
Section titled “More help”See check protection, troubleshooting and frequently asked questions. If the issue persists, contact your administrator without sharing sensitive data.