When to Verify AI Output

Always check these things in AI output:

High-risk areas (always verify):
- Medical, legal, or financial claims
- Statistics and data
- Citations and references
- Code logic and security

Medium-risk (spot check):
- Historical facts
- Technical explanations
- Contact information
- Dates and timelines

Lower-risk (but still review):
- General explanations
- Creative writing
- Formatting and structure

Remember: AI confidence does not equal accuracy. The most confident-sounding claims can be completely fabricated.