Record STMT-W4079CAVoluntary

Principles for the secure integration of Artificial Intelligence in Operational Technology

Guidelines International lawGlobalin force

This guidance outlines four key principles to help critical infrastructure owners and operators assess whether artificial intelligence (AI) is suitable for operational technology (OT) environments, understand the unique risks and impacts, establish clear rules and responsibilities, and embed safety and security into AI-enabled systems. It covers techniques ranging from traditional statistical modeling to machine learning and large language models, emphasizing human oversight and functional safety. The document is co-authored by international cybersecurity authorities including CISA, ASD's ACSC, and the NSA.

Summary and content fields are drawn from the document text and are pending editorial review.

Who issued it

Cybersecurity and Infrastructure Security AgencyGovernment · US

Lead issuer

Australian Cyber Security CentreGovernment · AU

Co-issuer

National Security AgencyGovernment · US

Co-issuer

Federal Bureau of InvestigationGovernment · US

Co-issuer

Canadian Centre for Cyber SecurityGovernment · CA

Co-issuer

Federal Office for Information SecurityGovernment · DE

Co-issuer

National Cyber Security Centre NetherlandsRegulator · NL

Co-issuer

National Cyber Security Centre New ZealandGovernment · NZ

Co-issuer

National Cyber Security Centre United KingdomRegulator · GB

Co-issuer

Legal forceVoluntary

Under whose authorityInternational law

Judged fromfull text

Who it governs

CompaniesPublic sector bodiesAI developers and providersAI deployers and users

Critical infrastructure owners and operators responsible for designing, deploying, and managing AI in operational technology environments should read this to safely integrate AI while maintaining system reliability.

Sectors

Defence and security · operational technology and critical infrastructureCross-sector

What the document contains

What's covered

The document covers principles and practical guidance for securely integrating machine learning, large language models, and AI agents into operational technology and critical infrastructure environments. It addresses risk assessment, data security, vendor transparency, governance frameworks, and safety/failsafe mechanisms aligned with the Purdue Model.

What's missing

It does not impose mandatory legal requirements, binding regulatory rules, or specific statutory enforcement penalties, as all guidance actions are explicitly voluntary.

Key takeaways

Integrating AI into operational technology (OT) environments introduces significant safety, security, and reliability risks that must be carefully managed.
Organizations should evaluate whether AI is truly the best solution for their operational needs compared to established alternatives.
Robust AI governance frameworks, supply chain transparency, and secure data handling are essential for protecting sensitive OT assets.
Human oversight, anomaly detection, and failsafe mechanisms must be maintained to ensure AI-enabled OT systems operate safely without causing physical harm.
Critical infrastructure operators should educate personnel on AI limitations and avoid relying on opaque models like large language models for direct safety-critical decisions.

Principles it states

Understand AI: Understand the unique risks and potential impacts of AI integration into OT environments, the importance of educating personnel on these risks, and the secure AI development lifecycle.Consider AI Use in the OT Domain: Assess the specific business case for AI use in OT environments and manage OT data security risks, the role of vendors, and the immediate and long-term challenges of AI integration.Establish AI Governance and Assurance Frameworks: Implement robust governance mechanisms, integrate AI into existing security frameworks, continuously test and evaluate AI models, and consider regulatory compliance.Embed Safety and Security Practices Into AI and AI-Enabled OT Systems: Implement oversight mechanisms to ensure the safe operation and cybersecurity of AI-enabled OT systems, maintain transparency, and integrate AI into incident response plans.

Outline

· Introduction
· Important Terminology
· Scope
· Types of AI Techniques
· AI Applications According to the Purdue Model
· Principle 1 – Understand AI
· Principle 2 – Consider AI Use in the OT Domain
· Principle 3 – Establish AI Governance and Assurance Frameworks
· Principle 4 – Embed Oversight and Failsafe Practices Into AI and AI-Enabled OT Systems
· Conclusion
· Resources
· Disclaimer
· Acknowledgements
· Appendix: Terminology
· References

IssuedNot recorded

Length8,424 words

FormatNot recorded

Scopeglobal

How it scores

Pending review

Each score is backed by a quote verified against the document text. Editorial review has not yet confirmed them.

Accountability

4/5

Clear responsibility for AI outcomes

Set clear rules, responsibilities and security measures to manage AI systems throughout their lifecycle.

At a glance

screening score · 2026-10-24

Fairness

0/5

Non-discrimination and equitable treatment

screening score · 2026-10-24

Human Dignity

0/5

Respect for the inherent worth of persons

screening score · 2026-10-24

Human Oversight

4/5

Humans remain in control of consequential decisions

Maintain human oversight and use monitoring and failsafe measures to ensure AI-enabled OT systems operate safely and securely.

At a glance

screening score · 2026-10-24

Privacy

4/5

Protection of personal data and private life

Protect proprietary and personal information within OT datasets, including by instituting protection from access abuse, intentional or inadvertent data poisoning, or dependency on synthetic, generated data.

2.2 Manage OT Data Security Risks for AI Systems

screening score · 2026-10-24

Transparency

4/5

Openness about how AI systems work and are used

Critical infrastructure owners and operators should demand transparency and security considerations from OT vendors regarding how AI technologies are embedded into their products.

2.3 Understanding the Role of OT Vendors in AI Integration

screening score · 2026-10-24

Enforcement

2/5

Consequences and compliance mechanisms with real force

Implementing regular audits and compliance testing to help identify potential issues and ensure ongoing adherence to AI governance requirements.

Principle 3 – Establish AI Governance and Assurance Frameworks

screening score · 2026-10-24

Impact Assessment

4/5

Required assessment of an AI system's risks/impacts before or during use

Assess whether AI is the best solution for your operational needs and understand the risks it may create for OT systems.

At a glance

screening score · 2026-10-24

Topics

Physical AI Governancecore

artificial_intelligence operational_technology critical_infrastructure cybersecurity secure_design risk_management