Security Resilience

CWE · MITRE source

CWE-1426Improper Validation of Generative AI Output

Abstraction: Base · CVEs in our corpus: 158

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.

Mapped by Security Resilience. This is an AI-native weakness the NVD catalog almost never assigns. We mapped 158 CVEs to it ourselves — a language-model classifier reads each CVE description and proposes the weakness, human-reviewed before it surfaces, with model poisoning and extraction cases deliberately routed to MITRE ATLAS rather than a synthetic CWE. The mappings are additive; NVD records are unchanged. Rows tagged CR below are ours.

Last updated: 15 October 2026 14:21 UTC

Cumulative inbound coverage

How completely the frameworks we cross-walk collectively cover this — the verdict is the strongest single mapping (overlapping partials are not summed); breadth shows the corroboration behind it.

Collective: partial · 1 mapping(s) from 1 framework(s): CSF 2.0 1 (partial)

See the full cumulative-coverage rollup →

Control responseHuman-reviewed

Answering this weakness across the control lifecycle, from our framework cross-walks.

Prevent
Stop it (NIST 800-53 / CSF Protect)
  • GV.PO-01
Detect
Catch it (CSF Detect / Respond)

—

Harden
Shrink the surface (DISA STIG)

—

Validate
Prove the fix (OWASP ASVS)

—

NIST 800-53 r5 controls PROPOSED for this weakness (0)AI-assisted

These are model PROPOSALS, not graded mappings. They come from controls_to_cwe_proposals, which carries no grade field at all; across the CWE pages 1,540 of 1,722 such associations (89%) have never been graded by anything. The heading said “address” until 2026-09-29, which asserted a relationship the data does not carry. Graded cross-walks are on the framework map, not here.

Control Title Family Why it addresses this CWE
No NIST controls proposed yet.

Top CVEs of this weakness type, ranked by Risk Priority

CVE Risk CVSS EPSS Published
CVE-2023-29374 CR9.19.80.39652023-04-05
CVE-2024-8309 CR8.39.80.13742024-10-29
CVE-2024-3271 CR7.89.80.02892024-04-16
CVE-2026-61447 CR7.810.00.02492026-07-11
CVE-2023-36188 CR7.79.80.01902023-07-06
CVE-2024-4343 CR7.79.80.02682024-11-14
CVE-2026-27966 CR7.79.80.02552026-02-26
CVE-2023-38896 CR7.69.80.01842023-08-15
CVE-2023-39659 CR7.69.80.01532023-08-15
CVE-2023-39661 CR7.69.80.01422023-08-15
CVE-2024-46946 CR7.69.80.01392024-09-19
CVE-2026-26015 CR7.69.80.01542026-04-29
CVE-2026-61539 CR7.610.00.01232026-08-21
CVE-2026-37003 CR7.69.80.01322026-08-27
CVE-2026-61732 CR7.610.00.01222026-09-24
CVE-2023-36095 CR7.59.80.01172023-08-05
CVE-2023-44467 CR7.59.80.01042023-10-09
CVE-2024-23752 CR7.59.80.01012024-01-22
CVE-2024-5826 CR7.59.80.00882024-06-27
CVE-2024-12366 CR7.59.80.01212025-02-11
CVE-2025-1497 CR7.59.80.01052025-03-10
CVE-2025-54795 CR7.59.80.01002025-08-05
CVE-2026-30741 CR7.59.80.01232026-03-11
CVE-2026-30306 CR7.59.80.00982026-03-30
CVE-2026-30308 CR7.59.80.00982026-03-30