CWE-1426 — Improper Validation of Generative AI Output
CWE-1426: Improper Validation of Generative AI Output
MITRE CWE weakness
| Kind | Weakness |
| Abstraction | Base |
| Status | Incomplete |
| Likelihood of exploit | — |
Description
The product invokes a generative AI/ML component whose behaviors and outputs cannot be directly controlled, but the product does not validate or insufficiently validates the outputs to ensure that they align with the intended security, content, or privacy policy.
Common consequences
- Integrity: Execute Unauthorized Code or Commands, Varies by Context
Mitigations
Architecture and Design — Since the output from a generative AI component (such as an LLM) cannot be trusted, ensure that it operates in an untrusted or non-privileged space.
Operation — Use "semantic comparators," which are mechanisms that provide semantic comparison to identify objects that might appear different but are semantically similar.
Operation — Use components that operate externally to the system to monitor the output and act as a moderator. These components are called different terms, such as supervisors or guardrails.
Build and Compilation — During model training, use an appropriate variety of good and bad examples to guide preferred outputs.
References
- CWE page: https://cwe.mitre.org/data/definitions/1426.html
- CWE list: https://cwe.mitre.org/data/index.html