# Security Level 5

> Creating optionality for frontier AI labs to reach SL5

Security Level 5 is a 501(c)(3) nonprofit developing the technical roadmap and standard for reaching SL5: machine, physical, network, personnel, and supply chain security capable of withstanding priority operations from the most capable nation-state actors.

## Organization Info

- Website: https://sl5.org
- Email: info@sl5.org
- Founded: June 2026

## What is Security Level 5?

SL5 is the fifth and highest security level defined in RAND's Securing AI Model Weights framework: machine, physical, network, personnel, and supply chain security capable of withstanding priority operations from the most capable nation-state actors. It is the level we are named after, and the one we exist to make reachable.

### Security Domains

- **Machine Security**: Protection of AI systems and hardware against unauthorized access and tampering.
- **Network Security**: Securing network infrastructure against sophisticated cyber threats.
- **Physical Security**: Securing physical devices against nation-state threats and compromised insiders.
- **Supply Chain Security**: End-to-end supply chain security and verification processes.
- **Personnel Security**: Comprehensive insider threat mitigation and personnel vetting.

## Core Projects

- [Novel Recommendations for SL5](https://sl5.org/projects/sl5-novel-recommendations): Security Level 5's practical blueprint for reaching SL5, compiling implementable controls, cost-benefit analyses, and pilot concepts organized as focused memos.
- [Sensitivity Levels Framework](https://sl5.org/projects/sensitivity-levels-framework): Industry-adapted personnel sensitivity levels for AI labs, establishing clear vetting and control criteria proportional to each role's access.
- [Security Level 5 Standard](https://sl5.org/sl5-standard): Risk management framework providing security controls and implementation guidance for achieving nation-state grade protection for frontier AI systems.
- [AI705](https://sl5.org/projects/ai705): A requirement-by-requirement analysis of whether ICD 705, the U.S. government's standard for secure classified facilities, would actually protect a frontier AI facility.
- [Project MoAT](https://sl5.org/projects/project-moat): An exploratory 1-2 MW mock SL5 datacenter build: a real test range for the SL5 Standard, AI705, and SenL before a production facility has to carry the learning curve.

## Security Reports

- [Machine Security - Novel Recommendations for SL5](https://sl5.org/reports/machine-security): Hardware and firmware security recommendations for achieving Security Level 5. Implementable controls for AI accelerators, verified boot, and tamper-proof infrastructure.
- [Personnel Security - Novel Recommendations for SL5](https://sl5.org/reports/personnel-security): Personnel security recommendations for achieving Security Level 5. Insider threat mitigation, AI agent governance, and industry-adapted vetting frameworks.
- [Network Security - Novel Recommendations for SL5](https://sl5.org/reports/network-security): Network security recommendations for frontier AI development. Inline network encryptors, accelerator interconnect encryption, air-gapped networks, and AI-enhanced cross-domain solutions.
- [Physical Security - Novel Recommendations for SL5](https://sl5.org/reports/physical-security): Physical security recommendations for frontier AI development. Two-person integrity, body-worn cameras, eliminating out-of-facility maintenance, side-channel defenses, and component verification.
- [Supply Chain Security - Novel Recommendations for SL5](https://sl5.org/reports/supply-chain-security): Supply chain security recommendations for frontier AI development. Architectural isolation, security-critical software review, hardware supply chain auditing, data integrity verification, and continuous red teaming.

## Tools

- [IL6 Control Catalog](https://sl5.org/projects/il6-control-catalog): An open-source, interactive catalog of security controls for Impact Level 6 systems. Browse, search, and understand FedRAMP High and DoD IL6 security requirements.
- [Secure Speech-to-Text](https://sl5.org/projects/secure-speech-to-text): A local, privacy-first toolkit for transcribing sensitive meetings and generating executive summaries, all without sending data to the cloud.
- [SenL Role-to-Level Calculator](https://sl5.org/tools/senl-role-to-level): Superseded by the Sensitivity Levels Framework. Interactive workshop tool for mapping organizational roles to Sensitivity Levels (SenL), kept available for reference.

## Key Documents

- [Novel Recommendations for SL5](https://sl5.org/pdfs/SL5_NOVEL-RECOMMENDATIONS.pdf)
- [Machine Security Report](https://sl5.org/pdfs/SL5_MACHINE-SECURITY-NOVEL-RECOMMENDATIONS.pdf)
- [Personnel Security Report](https://sl5.org/pdfs/SL5_PERSONNEL-SECURITY-NOVEL-RECOMMENDATIONS.pdf)
- [Network Security Report](https://sl5.org/pdfs/SL5_NETWORK-SECURITY-NOVEL-RECOMMENDATIONS.pdf)
- [Physical Security Report](https://sl5.org/pdfs/SL5_PHYSICAL-SECURITY-NOVEL-RECOMMENDATIONS.pdf)
- [Supply Chain Security Report](https://sl5.org/pdfs/SL5_SUPPLY-CHAIN-SECURITY-NOVEL-RECOMMENDATIONS.pdf)

## Contact

- Email: info@sl5.org
- Twitter: https://x.com/SL5TaskForce
- LinkedIn: https://www.linkedin.com/company/securitylevel5/