# Dual-Use AI Governance
**Entity class:** Security-governance concept
**Domain:** Artificial intelligence / biosecurity / cybersecurity
**Maturity:** Developed
## Definition
**Dual-Use AI Governance** manages capabilities whose legitimate scientific, medical or defensive uses are inseparable from cyber, biological or coercive misuse.
## Mechanism
The same model can discover a drug target or design a pathogen, find a vulnerability or exploit it. Governance therefore combines access tiers, monitoring, evaluation, incident response, hardware isolation and accountable exceptions rather than relying on a universal filter or unrestricted release.
## Article Context
[[articles/The Immediate AI Risk Is the Public|The Immediate AI Risk Is the Public]] develops this concept inside a paired argument: Bill Gates is right that frontier AI creates concrete cyber, biological and control dangers, while the administration is right that US/Israel must continue building at AI-superpower scale beside a permanent Chinese peer. The concept explains how those positions interact rather than forcing them into opposing camps.
[[articles/Vertically Integrating an AI Superpower from AI Factory to Citizen|Vertically Integrating an AI Superpower from AI Factory to Citizen]] supplies the material and strategic stack. [[articles/Safety Discourse Is Transition Discourse|Safety Discourse Is Transition Discourse]] supplies the historical transition and disclosure layer.
## Evidence Boundary
Dual use is not proof that every capability should be restricted. Controls must track demonstrated capability, preserve legitimate research and state their tradeoffs openly.
## Relationships
- [[wiki/Biosecurity|Biosecurity]]
- [[wiki/AI Safety|AI Safety]]
- [[wiki/AI Monitoring Substrate|AI Monitoring Substrate]]
- [[wiki/Control Problem|Control Problem]]
- [[wiki/AI Capability Diffusion|AI Capability Diffusion]]
- [[wiki/Compute Sovereignty|Compute Sovereignty]]
## Simple Reminders, Quotations, and Thoughts
> “When the capability to kill millions passes from nation-states to small groups, safety stops being a policy preference and becomes the price of civilization's continued existence.”
> **— Bryant McGill**, *The Immediate AI Risk Is the Public, 2026*
[[reminders/Existential Risk/Safety Is the Price of Civilizational Survival by Bryant McGill|Safety Is the Price of Civilizational Survival by Bryant McGill]]
> “A technology that its most experienced builders compare unfavorably to nuclear weapons cannot be governed by lawsuits filed after the funerals.”
> **— Bryant McGill**, *The Immediate AI Risk Is the Public, 2026*
[[reminders/AI Control/Lawsuits After Funerals Cannot Govern Artificial Intelligence by Bryant McGill|Lawsuits After Funerals Cannot Govern Artificial Intelligence by Bryant McGill]]
> “The first mass-casualty danger of artificial intelligence comes from a human hand holding a superhuman tool, long before any machine decides anything on its own.”
> **— Bryant McGill**, *The Immediate AI Risk Is the Public, 2026*
[[reminders/Existential Risk/First Mass-Casualty AI Danger Comes from a Human Hand by Bryant McGill|First Mass-Casualty AI Danger Comes from a Human Hand by Bryant McGill]]
> “A public that can only hear "safe" or "dangerous" will never govern a technology that is both, because it will spend every crisis choosing a team instead of understanding a mechanism.”
> **— Bryant McGill**, *The Immediate AI Risk Is the Public, 2026*
[[reminders/Risk Debate/Binary Publics Cannot Govern Dual-Use Technology by Bryant McGill|Binary Publics Cannot Govern Dual-Use Technology by Bryant McGill]]
> “A civilization that distributes the power of mass destruction to small groups while leaving its governance to an uninformed crowd has built a world in which luck is the only remaining safeguard.”
> **— Bryant McGill**, *The Immediate AI Risk Is the Public, 2026*
[[reminders/Existential Risk/Mass-Destruction Capability and Public Ignorance Leave Only Luck by Bryant McGill|Mass-Destruction Capability and Public Ignorance Leave Only Luck by Bryant McGill]]
> “We crossed the cyber threshold and we crossed the bio threshold early this year.”
> **— Bill Gates**, *The Ezra Klein Show, September 29, 2026*
[[reminders/Existential Risk/AI Crossed the Cyber and Bio Thresholds by Bill Gates|AI Crossed the Cyber and Bio Thresholds by Bill Gates]]
> “In other areas, do we just say, ‘Hey, release your drugs. There’s no FDA’? You can’t rely on the industry to self-regulate here. I mean, it’s just insane.”
> **— Bill Gates**, *The Ezra Klein Show, September 29, 2026*
[[reminders/AI Control/AI Industry Self-Regulation Is Insane by Bill Gates|AI Industry Self-Regulation Is Insane by Bill Gates]]
> “This makes nuclear weapons look like nothing.”
> **— Bill Gates**, *The Ezra Klein Show, September 29, 2026*
[[reminders/Existential Risk/AI Makes Nuclear Weapons Look Like Nothing by Bill Gates|AI Makes Nuclear Weapons Look Like Nothing by Bill Gates]]
> “The imminent risk is these are the most powerful tools ever.”
> **— Bill Gates**, *The Ezra Klein Show, September 29, 2026*
[[reminders/Existential Risk/The Imminent AI Risk Is the Most Powerful Tools Ever by Bill Gates|The Imminent AI Risk Is the Most Powerful Tools Ever by Bill Gates]]
> “[A serious user] either has to get a special permission copy or go use something where there are absolutely no safeguards at all.”
> **— Bill Gates**, *The Ezra Klein Show, September 29, 2026*
[[reminders/AI Control/Strict AI Safeguards Push Serious Users to Unmonitored Models by Bill Gates|Strict AI Safeguards Push Serious Users to Unmonitored Models by Bill Gates]]
## Sources / Provenance
- [[articles/The Immediate AI Risk Is the Public|The Immediate AI Risk Is the Public]] — canonical source for this node's role in public competence, AI safety, superpower competition and machine succession.
- [[articles/Vertically Integrating an AI Superpower from AI Factory to Citizen|Vertically Integrating an AI Superpower from AI Factory to Citizen]] — full-stack infrastructure and national-capability context.
- [[articles/Safety Discourse Is Transition Discourse|Safety Discourse Is Transition Discourse]] — institutional history, disclosure and transition context.