Intelligence Security Laboratories

Intelligence Security Laboratories (ISL) is a 501(c)(3) nonprofit research lab that describes its focus as "implementation-driven R&D for high-security AI systems" S-1706. It aims to build and demonstrate the security that critical AI deployments will need, and names tampering with or theft of a model by nation-state actors, other AIs or the model itself as the threats S-1706; see Model weights have not left the facility.

Its stated approach:

• ISL argues that studying each component in depth will not show that a data centre is secure as a whole system, so it aims to be the integrator for the whole problem S-1706.
• It applies STPA-Sec (Systems-Theoretic Process Analysis for Security), a system design method that emphasises unknown-unknown attack vectors, and uses formal methods wherever it can S-1706. For details of STPA-Sec, it points to RAND's secure inference data center design S-1706.
• It plans to hand the knowledge and technical artifacts it produces to other actors, and states that it does not expect to become the world's producer of secure compute S-1706.

Its executive director, Gabriel Kulp, is a co-author of the RAND report S-1510 S-1707. Before founding ISL, he was a fellow at RAND working on hardware-enabled governance mechanisms for GPU export controls and on international verification of agreements S-1707. Tom Gardiner is ISL's staff director and Paul Murley its technical director S-1707.