How to meet it
OWASP AISVS v1.01
Model Behavior, Output Control & Safety Assurance · Hallucination Detection & Mitigation
Estimate confidence in answers
Verify that the system assesses the reliability of generated answers using a confidence estimation method.
OWASP AISVS v1.01, requirement 7.2.1
In plain English
Estimate how reliable each answer is, so the system can treat shaky answers differently.
What good looks like
- A confidence method is chosen for the use case and documented.
- Confidence scores are stored with responses.
- The method is checked against real error rates.
How teams usually do it
- Checking the answer against retrieved sources.
- Sampling several answers and measuring agreement.
- Token probabilities where your model API exposes them.
Evidence an auditor would accept
- The chosen method and calibration results.
Common mistakes
- Asking the model how confident it is and trusting the reply.
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.