How to meet it
OWASP AISVS v1.01
Input Validation · Content & Policy Screening
Check screening works in other languages
Verify that prompt content classification is evaluated for unsupported languages.
OWASP AISVS v1.01, requirement 2.2.2
In plain English
Check how well your content screening works in languages you do not officially support, because attackers switch language to get past filters.
What good looks like
- Screening is tested in several languages beyond the supported ones.
- Inputs in languages the classifier handles poorly are blocked or sent for stricter handling.
- Results are reviewed when the classifier changes.
How teams usually do it
- Language detection in front of the classifier, with a policy for unsupported languages.
- A multilingual test set covering your harmful content categories.
Evidence an auditor would accept
- Test results by language.
- The policy for unsupported languages.
Common mistakes
- Testing only in English and assuming the classifier is equally good elsewhere.
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.