# Mutual Assured AI Malfunction **Entity class:** Strategic AI-denial doctrine **Domain:** AI security / deterrence / strategic competition **Doc Type:** Developed Concept Node **Maturity:** Proposed **As of:** 2026-09-20 ## Definition **Mutual Assured AI Malfunction** is a proposed deterrence framework in which rival powers maintain capabilities to disrupt, corrupt, or disable the infrastructure supporting advanced AI systems. It appears in the 2025 _Superintelligence Strategy_ associated with Dan Hendrycks, Eric Schmidt, and Alexandr Wang. ## Evidence Ledger - **Established:** The 2025 strategy publicly proposed AI-denial and mutual-vulnerability mechanisms. - **Strongly indicated:** AI infrastructure is being theorized through deterrence concepts previously applied to nuclear and cyber systems. - **Plausible:** States are developing operational capabilities aligned with the doctrine under other program names. - **Unresolved:** Adoption as national policy. Promotion requires doctrine, budget, command guidance, or operational planning records; demotion requires authoritative policy rejecting the framework. ## Sources / Provenance - Hendrycks, Schmidt, and Wang, [_Superintelligence Strategy_ (2025)](https://www.nationalsecurity.ai/). - [[articles/A History of Machine Intelligence|A History of Machine Intelligence]]. ## Relationships - **Belongs to:** [[wiki/Denial Wars|Denial Wars]]. - **Extends:** [[wiki/Broken Nest|Broken Nest]]-style denial from fabs to AI systems and dependencies.