How to meet it
OWASP AISVS v1.01
Access Control & Identity for AI Components & Users · AI Resource Authorization & Classification
Carry data labels through to outputs
Verify that data classification labels propagate to downstream resources (embeddings, prompt caches, model outputs).
OWASP AISVS v1.01, requirement 5.2.7
In plain English
When data is labelled confidential, that label should follow it into embeddings, caches and model outputs, so later controls know what they are handling.
What good looks like
- Embeddings carry the classification of their source.
- Prompt and response caches record the highest classification involved.
- Outputs inherit classification for logging and display.
How teams usually do it
- Classification fields in vector metadata, set from the source at ingestion.
- Logging and caching code that reads and applies the label.
Evidence an auditor would accept
- Metadata schema and ingestion code.
- Samples showing labels on derived data.
Common mistakes
- Labelling source documents but treating their embeddings as unclassified.
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.