ISO 27090 and the NIST AI Risk Management Framework both help organisations manage AI risk, but at different widths. ISO/IEC 27090:2026 is focused on the security threats specific to AI and how to mitigate them. The NIST AI RMF is a broader, voluntary framework for managing AI risk across trustworthiness, bias, safety and security. They complement each other rather than compete.
What is the difference in scope?
The NIST AI RMF is wide, ISO 27090 is deep on security. The NIST AI Risk Management Framework, published by the US National Institute of Standards and Technology in January 2023, helps organisations manage the full range of AI risks, organised around four functions, Govern, Map, Measure and Manage, and a set of trustworthiness characteristics such as validity, safety, fairness, accountability and security. Security is one strand within it. ISO 27090 does one thing in detail: it describes the security threats and compromises specific to AI systems, such as data poisoning, model theft and prompt injection, and how to detect and treat them. One frames the whole AI risk conversation; the other is the security specialist inside it.
What is the difference in status?
Both are voluntary, but they come from different places. The NIST AI RMF is a US government framework, voluntary by design, widely referenced internationally. ISO 27090 is an international standard from ISO and IEC, and although it is guidance rather than a certifiable requirements standard, it carries the weight of the same committee behind ISO 27001. Neither is law in Australia, and neither can be certified. The practical difference is familiarity: organisations already using NIST frameworks will slot the AI RMF in easily, while organisations built around the ISO 27000 series will find ISO 27090 a more natural fit.
| NIST AI RMF | ISO 27090 | |
|---|---|---|
| Author | NIST (US) | ISO and IEC |
| Scope | All AI risk (trust, bias, safety, security) | AI specific security threats |
| Structure | Govern, Map, Measure, Manage | Security practice, threats, mitigations |
| Status | Voluntary framework | Guidance standard, not certifiable |
| Published | January 2023 | October 2026 |
Can you use both together?
Yes, and that is the sensible approach. Use the NIST AI RMF to organise how the business governs and manages AI risk end to end, then use ISO 27090 as the detailed source for the security threats the Map and Measure steps need to identify and assess. Where an organisation also runs ISO 42001, the AI management system standard, the three line up neatly: 42001 is the certifiable governance system, the NIST AI RMF is a complementary risk method, and ISO 27090 is the AI security threat library all of them can draw on. None of the three replaces the others.
Which should an Australian organisation pick?
Pick by what you already run, then add ISO 27090 for the security detail. If your organisation uses ISO standards and is pursuing ISO 42001, lead with that and treat ISO 27090 as the AI threat reference. If you use NIST frameworks such as the NIST CSF, the AI RMF will feel natural, and ISO 27090 still adds value as the security specific layer. For most Australian organisations the honest answer is that these are not either or choices: the frameworks organise the work, and ISO 27090 makes the AI security part concrete. Cybernion helps organisations choose and apply the right mix through AI governance readiness.
Frequently asked questions
Is the NIST AI RMF a security standard?
Not only. It is a voluntary framework for managing AI risk broadly, including trustworthiness, bias and safety, of which security is one part. ISO 27090 is focused on AI security threats.
Are either of them mandatory?
No. The NIST AI RMF is voluntary, and ISO 27090 is guidance. Neither is law in Australia.
Can you use both?
Yes. The NIST AI RMF organises overall AI risk management; ISO 27090 supplies the AI specific security threats and mitigations to feed into it.
Written by Gaurav Vikash, an ASD endorsed IRAP assessor and senior cyber security leader with 18 years of experience across Australia, the UK and Asia, including CISO and senior security leadership roles. He holds CISSP, CISA, CISM and CRISC and is an ISO 27001 Lead Auditor and ISO 42001 Lead Implementer, and speaks regularly at industry conferences.
If you are choosing an AI risk approach, talk to us.
Sources:
- ISO/IEC 27090:2026, Cybersecurity, Artificial intelligence, Addressing security threats and compromises to AI systems, October 2026
- NIST AI Risk Management Framework (AI RMF 1.0), January 2023
- ISO/IEC 42001:2023, Information technology, Artificial intelligence, Management system, December 2023
Last updated: 24 October, 2026
