AI Component

Training Data

Training data is both the model's most valuable input and its most underprotected one. Three problem classes dominate. First, poisoning: an attacker who can influence a public dataset, a web crawl, or a fine-tuning corpus can plant backdoors or biases that survive into the deployed model — BadNets-style attacks on image classifiers, trigger-phrase attacks on LLMs, and reward-hacking on RLHF datasets. Second, memorization and leakage: models can regurgitate verbatim training data, exposing PII and copyrighted content; this has driven the active New York Times v. OpenAI litigation and is a recurring GDPR concern. Third, provenance: when training data origins are unclear, downstream users inherit legal and security risk they can't assess. EU AI Act Article 10 (Data Governance) and ISO 42001 Annex A treat training-data quality as a controlled asset. Defenses: data lineage tracking, deduplication, PII scrubbing before training, and adversarial training against known trigger families.

287
Total CVEs
15
Pages
Page 10 of 15
Current
SeverityCVECVSS
HIGHCVE-2026-118168.1
CRITICALCVE-2026-48797-
CRITICALCVE-2025-713219.8
HIGHCVE-2025-713768.1
HIGHCVE-2026-52812-
CRITICALCVE-2026-564459.1
HIGHCVE-2026-581168.8
HIGHCVE-2025-713508.1
MEDIUMCVE-2026-124805.5
HIGHCVE-2026-575168.8
UNKNOWNCVE-2026-8387-
CRITICALCVE-2026-235379.1
HIGHCVE-2026-562097.1
CRITICALCVE-2026-145359.8
MEDIUMCVE-2026-554385.8
HIGHCVE-2026-550757.4
HIGHCVE-2026-489578.8
CRITICALCVE-2026-616679.9
LOWCVE-2026-156053.1
UNKNOWNCVE-2026-45535-

Page 10 of 15