SL5 Standard for AI Security
A NIST SP 800-53 overlay for frontier AI infrastructure achieving nation-state-level security. 43 controls across 10 families by the Security Level 5 Task Force.
Version 0.1 - Long lead times An integrated security control profile for frontier AI infrastructure, anchored in NIST SP 800-53 and extended with supply-chain and facility requirements, targeting nation-state-level security by 2028/2029. This version 0.1 focuses on long lead time controls, prioritizing measures that must be planned years in advance. Future versions will extend to shorter-lead-time measures and fine-grained controls. 60 controls across 12 families - March 2026 (updated June 23, 2026) About This Document Security Level 5 (SL5) is a security posture for AI systems that…
saved by
related reading
- The case for ensuring that powerful AIs are controlled — LessWronglesswrong.com
- Statement on the US government directive to suspend access to Fable 5 and Mythos 5 \ Anthropicanthropic.com
- 2028: Two scenarios for global AI leadership \ Anthropicanthropic.com
- Project MoAT - Mock SL5 Datacenter Buildsl5.org
- America’s AI Action Planwhitehouse.gov
- Promoting Advanced Artificial Intelligence Innovation and Security – The White Housewhitehouse.gov
- Off Target | CNAScnas.org
- Frontier safety blueprintcdn.openai.com
- GPT-5.5 and the broken state of government evalstransformernews.ai
- An Action Plan for American Leadership in AI | IFPifp.org
- What Should Be Donehyperdimensional.co
- Soft Nationalization: how the USG will control AI labs — LessWronglesswrong.com