International

ISO/IEC 27090:2026

78 controls. 0 other frameworks in our corpus share controls with it. Here is all of it, and how much of it you are already doing.

Page built . This page is derived from the framework corpus, which changes when the corpus is extended rather than daily.

78 controls 0 frameworks share controls with it International verified against its source document

Every control below is one this framework asks for. The right hand column counts how many other frameworks in our corpus carry the same control, which is the difference between doing this work once and doing it again for the next standard.

There is no implementation kit for this framework yet. The control list and the overlap above are free and complete.

What you already have

Frameworks whose controls overlap this one, most first. If you run any of them, the count is roughly what you have already evidenced.

No measured overlap with another framework in the corpus.

Every control

CodeControlAlso in
1Scope of the document0
2Normative references of the standard0
4Abbreviated terms and acronyms0
5.1Security controls applied to AI systems0
5.2Establishing an AI governance programme0
5.3Zero trust principles for AI systems0
5.4Software engineering practices for AI0
5.5.1Purpose of supply chain management0
5.5.2AI supply chain threat assessment0
5.5.3Complexity of AI supply chains0
5.5.4Supply chain attack vectors0
5.5.5AI supply chain security measures0
5.6Threat modelling for AI systems0
5.7Red teaming of AI systems0
6.1Types of AI system compromise0
6.10.1Output injection attacks description0
6.10.2Failures from output injection attacks0
6.10.3Detect output injection attacks0
6.10.4Mitigate output injection attacks0
6.11.1General recommended practices pointer0
6.11.2.1Limiting data to reduce impact0
6.11.2.2Data minimization0
6.11.2.3Minimal retention0
6.11.2.4Data encryption and obfuscation0
6.11.2.5Limit exposure of AI technical details0
6.11.3.1Determine cause of unwanted model behaviour0
6.11.3.2Human oversight and guardrails0
6.11.3.3Least privilege for AI agents0
6.11.3.4Tell stakeholders how the system works0
6.11.3.5Continuous validation against test data0
6.11.3.6Explain individual system decisions0
6.11.3.7Test for unwanted bias0
6.11.4Provenance for AI systems and data0
6.2.1How data poisoning works0
6.2.2Failures caused by data poisoning0
6.2.3Detect poisoned training data0
6.2.4Mitigate data poisoning attacks0
6.3.1Evasion attack description0
6.3.2Failures from evasion attacks0
6.3.3Detect evasion attacks0
6.3.4Mitigate evasion attacks0
6.4.1Membership inference description0
6.4.2Failures from membership inference0
6.4.3Detect membership inference0
6.4.4Mitigate membership inference0
6.5.1Model exfiltration description0
6.5.2Failures from model exfiltration0
6.5.3Detect model exfiltration0
6.5.4Mitigate model exfiltration0
6.6.1Model inversion description0
6.6.2Failures from model inversion0
6.6.3Detect model inversion0
6.6.4Mitigate model inversion0
6.7.1Direct model poisoning description0
6.7.2Failures from direct model poisoning0
6.7.3Detect direct model poisoning0
6.7.4Mitigate direct model poisoning0
6.8.1Sensitive model output description0
6.8.2Failures from sensitive model output0
6.8.3Detect sensitive model output0
6.8.4Mitigate sensitive model output0
6.9.1Prompt injection description0
6.9.2Failures from prompt injection0
6.9.3Detect prompt injection0
6.9.4Mitigate prompt injection0
7.1Assess mitigations as a system0
7.2Check for conflicting mitigation effects0
7.3Keep mitigations consistent across the life cycle0
7.4Monitor for mitigation deterioration0
7.5Log and monitor model use0
7.6Protect the development environment0
7.7.1Detect out of distribution inputs0
7.7.2Detect adversarial inputs0
7.8Throttle model use per user0
7.9Limit model size and memorisation0
Annex APair threat mappings with conventional controls0
Annex BProtect training data in development and operation0
Annex CFederated learning risks and protections0

Tell me when ISO/IEC 27090:2026 files something new

One email when a public company newly discloses something this framework governs, naming the company and what our corpus says it puts in scope. Nothing else, and one click to stop.

What this page is

A control-level reference for ISO/IEC 27090:2026, drawn from our framework corpus. Control codes and titles are references to the standard, not reproductions of it. The overlap counts and the auditor artefacts are our own work and are the part you will not find elsewhere.

Measure this against what you already run · All frameworks · Today's edition