One-sentence summary
ISO/IEC 27090:2026 addresses security threats to AI systems and responses when compromised, already published; ISO/IEC 27091 covers AI privacy protection at the Final Draft (FDIS) stage — one for security, one for privacy, used together.
Standards
- ISO/IEC 27090:2026: Cybersecurity — AI — Addressing security threats and compromises to AI systems; published October 2026 (Edition 1).
- ISO/IEC FDIS 27091: Cybersecurity and Privacy — AI — Privacy protection; Final Draft International Standard.
Key Parameters
| Dimension | ISO/IEC 27090:2026 | ISO/IEC FDIS 27091 |
|---|---|---|
| Topic | Security threats and compromises to AI systems | Privacy protection in AI |
| Status | Published (2026-10, Edition 1) | FDIS final draft stage |
| Area | ICS 35.030 (IT security) | Cybersecurity and privacy |
| Price | CHF 204 | Not announced |
| Focus | Threat detection, intrusion, model tampering | Personal data processing and privacy risk |
Key Differences
- Angle: 27090 answers how to respond when AI systems are attacked, breached or tampered; 27091 answers how to protect privacy when AI processes personal data.
- Maturity: 27090 is a citable published international standard; 27091 is still FDIS — enterprises can align early but should track the final text.
- Use case: Security engineering and red teams can cite 27090; data-protection compliance and privacy impact assessments can follow the 27091 direction.
Guidance (Compliance View)
- AI system security: use 27090 as the baseline for threats and response.
- Data protection and privacy assessment: track 27091 final and pre-assess against the FDIS framework.
- Use both in parallel to cover the security and privacy compliance lines.
Sources
- ISO: ISO/IEC 27090:2026
- ISO: ISO/IEC FDIS 27091