A safety filter blocks some useful content and misses some harmful content.
Tune it for the product's risk.
Choose two practices.
Measure false positives and false negatives on representative labeled cases, select thresholds with risk owners, and provide review or fallback paths.
Detailed explanation
Both overblocking and missed harmful content are visible.
The operating point reflects business and safety risk.
Uncertainty and novel cases remain.
Useful content may be blocked and users may bypass controls.
Check official documentation
Compare the following points with the official documentation.
AWS公式AIF-C01 Domain 2.3のGuardrails、コンテンツフィルター、安全性評価を確認する。
Official references
What to confirm
安全性と有用性をフィルターの閾値・誤り・版管理で運用できる。
Key points
- Threshold
- False positive
- False negative
Notes
- Environment: AWS公式AIF-C01試験ガイドとAWS公式ドキュメントの確認
- Compare the definition with the AIF-C01 objective and the service documentation before applying it.
- Keep the concept separate from a particular answer choice or implementation step.
Foundation review
Content filter
A control that detects, blocks, masks, or transforms harmful, prohibited, or sensitive content.
False positive
Safe content incorrectly classified as harmful or blocked.