Content Security Protection System
Pocket AI has a built-in triple protection mechanism: 1. localized sensitive word filtering library (automatically updated weekly); 2. conversation content sandbox isolation; 3. configurable content security level (Settings-Security Policy). Users can additionally take: add customized blocking words, turn on "Teen Mode", and regularly clean up sensitive conversation records (right-click menu - Security Erase).
- Deep Protection: Replace the filter rule files in the models/safety subdirectory
- Emergency Switch: Input "Emergency Stop" 3 times continuously to terminate the current session immediately.
- Audit function: Enable "Dialog Log" to trace back all interactions.
This answer comes from the articlePocket AI: offline AI assistant running in your phone, adapted for DeepSeek-R1 (5.37GB)The
May not be reproduced without permission:AI productivity tools " How to prevent offline AI assistants from generating unsafe conversational content?
