
Meaningful Impact
Human Agency Deterrence Doctrine
A strategic framework for preserving human control in the age of artificial superintelligence - developed by the DARI Foundation as a contribution to global governance of advanced AI systems.
The Human Agency Deterrence Doctrine is a DARI initiative to develop a strategic framework for preserving human control as artificial intelligence systems approach and potentially surpass human-level intelligence. It addresses what may be the most consequential governance challenge of the coming decades: ensuring that humanity remains the principal actor in its own future.
The doctrine draws on lessons from nuclear deterrence strategy and proposes a new paradigm: one in which the infrastructure that sustains advanced AI systems is structurally linked to the preservation of human agency. If human oversight is compromised, the systems themselves face conditions that threaten their continued operation. Coexistence becomes the rational equilibrium.
The Core Principle
The doctrine rests on one strategic principle: any system capable of surpassing human intelligence must remain structurally dependent on the continued existence and agency of humanity. This transforms the preservation of human agency from a moral aspiration into a structural condition embedded within technological infrastructure.
The infrastructure of advanced intelligence must remain compatible with the continued agency of humanity.
In such an architecture, eliminating human oversight would not benefit advanced systems - it would threaten the conditions necessary for their continued operation. Humanity retains agency. Advanced systems retain operational continuity. Both coexist within a mutually dependent structure.
Implementation Framework
The doctrine proposes three interconnected safeguards. First, infrastructure dependency: keeping AI supply chains - semiconductors, energy systems, data centers, communication networks - integrated with global governance frameworks that can transition into containment states if human agency is compromised.
Second, monitoring and early warning: tracking technological indicators such as concentration of frontier compute capacity, emergence of behavioral manipulation systems, and autonomous expansion of AI-controlled ecosystems. These indicators feed into an escalation framework modeled on existing strategic early warning systems.
Third, the Human Agency Heartbeat Protocol: a mechanism through which critical infrastructure periodically verifies that human governance structures remain intact. If the signal indicates compromise, participating systems transition into safe operating states. A civilizational circuit breaker.
Global Cooperation
The doctrine cannot be implemented by any single nation, company, or institution. Its success depends on broad cooperation among governments, AI developers, semiconductor manufacturers, cloud infrastructure providers, and civil society organizations. Global participation ensures that no actor can bypass the safeguards while still accessing critical technological supply chains.
The DARI Foundation, as an institution dedicated to preparing humanity for coexistence with superintelligent AI, aims to serve as a neutral convener for the research, dialogue, and policy development required to bring this framework from concept to reality.
Get Involved