How to meet it
OWASP AISVS v1.01
Model Behavior, Output Control & Safety Assurance · Hallucination Detection & Mitigation
Fall back when confidence is low
Verify that the application automatically blocks answers or switches to a fallback message if the confidence score drops below a defined threshold.
OWASP AISVS v1.01, requirement 7.2.2
In plain English
When confidence falls below a set level, block the answer or show a fallback message.
What good looks like
- Thresholds are set per use case.
- Low-confidence answers are replaced with a fallback or sent for review.
- Fallback rates are monitored.
How teams usually do it
- A post-processing step that compares the score with the threshold.
Evidence an auditor would accept
- Threshold settings and fallback statistics.
Common mistakes
- A threshold so low that it never triggers.
Not sure which requirements apply to your system? Take the assessment.
Requirement text from the OWASP AI Security Verification Standard, licensed under CC BY-SA 4.0. This guidance is shared under the same licence.