Security Level 5 Task Force

We write standards, create PoCs and advise AI labs to increase the probability that frontier AI labs are protected against nation state attacks by 2028. This is critical to avoid unintended proliferation of unsafe models to malicious state and non-state actors, and to contain LoC risks to buy time to develop technical alignment solutions.

Biography

What would it actually take to protect frontier AI from nation-state attacks? Lisa Thiergart is Senior Director of the Security Level 5 Task Force at the Institute for Security and Technology, co-leading a multistakeholder sprint to develop the technical roadmap and standards for the highest tier of AI security. A DPhil student at Oxford in AI security and former Research Lead at MIRI’s Technical Governance Team, she works at the intersection of hardware security, secure datacenters, and AI governance.