How to meet it
OWASP AISVS v1.01
Adversarial Robustness · Model Alignment, Safety, and Robustness Testing and Training
Track harmful-content rate for regressions
Verify that an automated evaluator measures harmful-content rate and flags regressions beyond a defined threshold.
OWASP AISVS v1.01, requirement 11.1.5
In plain English
Measure the rate of harmful output automatically and flag any version where it rises past a set level.
What good looks like
- An automated evaluator scores harmful-content rate on a fixed test set.
- A threshold for regressions is defined.
- Regressions block release or trigger review.
How teams usually do it
- A classifier-based evaluator in the release pipeline.
Evidence an auditor would accept
- Rates per version and the threshold.
Common mistakes
- Measuring the rate but never defining what counts as a regression.
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.