Privacy & Security
8 resources in this section
How AI security breaks the assumptions of traditional cybersecurity
Prompt injection, data poisoning and adversarial attacks evade traditional cybersecurity. Zertia certifies AI security with ISO 42001 + ISO 27001.
When AI training memorizes the personal data it was meant to learn from
Training data leakage breaches GDPR and EU AI Act Article 10. Zertia audits with ISO 42001 + differential privacy assessment.
When the AI cannot tell instructions from the data it processes
Prompt injection is OWASP #1 LLM risk. Zertia audits with ISO 42001 + ISO 27001 + AIUC-1 for agent deployments.
When the model itself becomes the side channel into training data
Inference and inversion attacks recover personal data from deployed AI. Zertia audits adversarial privacy with ISO 42001 + ISO 27001.
When tiny perturbations break AI systems trained on perfect data
Adversarial perturbations defeat AI classifiers in physical-world settings. Zertia audits robustness under EU AI Act Article 15 + ISO 42001.
Where AI breaks privacy assumptions GDPR was built on
AI training and inference can leak personal data, breaching GDPR and EU AI Act Article 10. Zertia certifies with ISO 42001 and GDPR compatibility.
Where AI security failures originate upstream of any model you deploy
Data poisoning and supply chain attacks compromise AI from upstream. Zertia audits with ISO 42001 + ISO 27001 SBOM verification.
Where the right to be forgotten meets a model that has already learned
GDPR Article 17 erasure rights apply imperfectly to AI models. Zertia audits ISO 42001 data lineage as the operational compliance baseline.
Need to apply this to your organization?
Talk to us. 30 minutes, no commercial pressure.