AI Resource
What Should Prevention Organizations Avoid Entering Into AI Tools?
Prevention organizations should define what information staff may not enter into AI tools before AI use becomes routine.
When in doubt, staff should not enter sensitive, identifiable, confidential, or protected information into an AI tool without an approved process.
Key points
- Avoid personally identifiable information.
- Avoid protected health, education, or client information.
- Avoid confidential partner or agency details.
- Avoid sensitive incident details or unpublished internal documents.
A practical approach
- Create a prohibited-input list.
- Train staff on examples.
- Use de-identified or synthetic examples when appropriate.
- Require approval for higher-risk use cases.
How GS Federal can help
GS Federal can help organizations create privacy-aware AI guidance and staff training.