UK publishes AI Cyber Security Code of Practice with 13 principles, later standardized as ETSI TS 104 223

Key facts

As stated in the sources, with where to find them.

• 13 principles across five phases: secure design (4), secure development (5), secure deployment (1), secure maintenance (2), secure end of life (1).Code structure
• Provision 2.3 requires an audit trail covering operation and lifecycle management of models, datasets and prompts; provision 2.6 limits permissions to those required for functionality.Principles 2 (audit trail, access)
• Principle 12 asks operators to monitor system behaviour, including internal states where useful, to detect anomalies and unexpected behaviour over time.Principle 12
• ETSI TS 104 223 (April 23, 2025) expands the 13 core principles into 72 trackable principles and lists indirect prompt injection among covered threats.ETSI press release

Findings that cite this record

No tracked finding cites this record yet.

Key questions this bears on

• Can prompt injection against AI agents be reliably defended?Not reliably. Defenses reduce injection but none has eliminated it; limiting what untrusted input can trigger is the most defensible approach.
• Can AI agents defend and oversee systems on their own?Not yet. Agents are weak on realistic defensive benchmarks and monitors can be evaded; assistants help analysts who stay in charge.

Sources

• Code of Practice for the Cyber Security of AIGOV.UKJan 31, 2025Primary sourceAccessed Sep 25, 2026
• ETSI Technical Specification sets international benchmark for securing Artificial IntelligenceETSIApr 23, 2025Primary sourceAccessed Sep 25, 2026

Related records