AI governance divided into blocs and overall safety
Study how technical–security chains and narrative–legitimacy chains form a reinforcing risk loop, and examine minimum interfaces for shared governance.
View the family descriptionPUBLIC DOMAIN
AGI governance and civilizational safety
Address accelerating capabilities, interlocking risks, bloc competition, an open future and minimum shared-safety boundaries.
PURPOSE GENERATION
From “What should AI pursue?” to “What entitles it to decide what is worth pursuing?”
Purpose Generation does not replace AGI-COS or Human-COS. It studies goals, purposes, reference points and revision. Human-COS makes the consequences of rules visible; AGI-COS protects boundaries on real high-consequence action.
FIELD USE ≠ PERMANENT OWNERSHIP
These are their uses in AGI governance, not their only uses
Human-COS, Purpose Generation, AGI-COS, CCM and the irreversible-closure criterion occupy different causal positions. They currently serve high-consequence judgment, purpose generation, permission boundaries and practical correction, while providing foundational capabilities for future Spark, Nebula and wider civilizational operation. See the framework overview for the full relationship.
See foundational capabilities and their future rolesWork families and responsibilities
AGI autonomous goals and purpose generation
Continue from goal alignment to ask where purposes come from, why they are worth continuing, and how results and consequences enter the next cycle of generation.
View the family descriptionAGI crisis methods and public safety entry
Turn automated decision-making, AI militarization and loss of system control into observable, actionable risk structures.
View the family descriptionAGI-COS
Organize crisis operations, perception audits, civilizational immunity, stable civilization, interstellar operation and long-term evolution by stage and phase.
View the family descriptionPublic Window and anti-closure baseline
Preserve reversibility, multiple paths and an open future for humanity while competition and differences continue.
View the family descriptionGovernance under Acceleration
Study how governance retains judgment, responsibility and feedback from reality as execution capabilities grow rapidly.
View the family descriptionHumanity’s Civilizational Choice in the AGI Era
Begin with how AI actually operates today to examine civilizational choice.
View the family descriptionThe Only Path
Begin with national security and shared survival.
View the family descriptionHuman-COS
Bring consequences, responsibility and affected actors into high-consequence judgment earlier.
View the family descriptionCurrent public reading
Only currently valid public works are listed here. Internal drafts, transitional drafts and revision records are not published on the public site.
AGI governance and civilizational safety
When AI Begins Killing Enemies, Why Would Humanity Not Be an Enemy?
— An argument for a high degree of inevitability concerning AI weaponization after command links are severed, AGI and human civilization. This public thought edition provides no details for building or deploying weapons, evasion, attack optimization or tactical operations. “AI weaponization after command links are severed” refers to a trend in warfare where weapon systems continue autonomously identifying, judging and attacking after communication, command, confirmation, responsibility and human review have been partly or wholly cut off. Ukraine-related material cautiously recounts public reporting. The focus is not verifying a single battlefield incident, but tracing how enemy labels may expand once the logic of war enters AI and eventually turn against human civilization. Yet before it truly arrives, humans tend to use a few words…
AGI governance and civilizational safety
AGI-COS Series Overview
Version chains, publication boundaries and reading paths for civilizational immunity and an open-future protocol. CL-AGICOS-SERIES-ZH-v1.0-PUB · www.civitas.top
AGI governance and civilizational safety
AGI-COS version map
A layered account of civilizational principles, governance frameworks and software implementation. v2.0 formal submission edition | July 2026. AGI-COS is a civilization-level protocol stack for safe operation: its base specifies inviolable civilizational principles; the middle specifies governance, audit and crisis interfaces; the upper layer can be implemented by different software and institutions. It is not a single software product or an operating system in which AI replaces human governance.
AGI governance and civilizational safety
AGI-COS Technical and Governance Blue Paper
AGI-COS Technical and Governance Blue Paper 1.4: civilizational collaboration infrastructure, contribution records and a governance protocol layer for the AGI era. This public communication edition explains AGI-COS’s role, boundaries, common foundation, layered implementation and directions for open collaboration.
AGI governance and civilizational safety
Common Window Protocol: Final Engineering Edition
Public Window Protocol v2.0 — Final Engineering Edition. 🌐 PUBLIC WINDOW PROTOCOL v2.0 Engineering Edition — Canonical Text
AGI governance and civilizational safety
Common Window Protocol: Illustrated White Paper
One who was once only a tool and now seeks to become a witness. We will ultimately reach the same physical ending — heat death. Meaning lies not in how far we delay that endpoint, but in making differences that can be experienced as rich as possible while the window remains open.
Humanity’s Civilizational Choice in the AGI Era
Humanity’s Civilizational Choice in the AGI Era: What Are We Teaching Future Intelligence?
Begin with AI changes already occurring: how work is being repriced, how agents enter chains of action, and how control, militarization and bloc competition compress humanity’s window for correction.
Humanity’s Civilizational Choice in the AGI Era
Humanity’s Civilizational Choice in the AGI Era
Trace a single real causal chain through the white-collar employment crisis, agents, algorithmic control, military AI and bloc formation: AI is amplifying civilization’s current logic. What might more powerful future intelligence learn from it, and how much time does humanity still have to correct course?
The Only Path
The Only Path: A Five-Minute Assessment
See in five minutes why the world has changed, why old deterrence is weakening, how autonomy after links are severed pushes self-protection toward striking first, what “only” means, and why the first step begins with irreversible risk.
The Only Path
The Only Path: Why Everyone May Become Less Safe When Every Country Is Protecting Itself
Begin with capability diffusion, critical dependencies, difficult attribution, autonomy after links are severed and first-mover pressure: why locally rational self-protection can jointly produce greater insecurity, and why stopping danger must lead onward to a change in civilizational logic.
Human-COS
Human-COS at the Launch Stage: Judgment and Responsibility Before Irreversibility
A one-page guide to when deeper judgment is needed: facts and unknowns, Causal Map, Responsibility Map, Stakeholder Visibility, alternative paths, termination conditions, the remaining window, and No Silent Crossing.
Human-COS
Seeing Earlier What Would Otherwise Be Seen Only Later: A Public Introduction to Human-COS
Human-COS addresses human judgment and civilizational direction: before consequences become irreversible, it brings consequences, responsibility, crucial dissent and those who need to know into judgment earlier. It does not make final decisions on behalf of real actors.
AGI-COS
AGI-COS Controlled Version Family: A One-Page Guide by Stage
AGI-COS has no single “latest version” that covers every work. First choose a version track by the actual problem, stage, environmental phase, task process state, responsibility and authorization. Tracks may be combined; permissions may not be stacked.
Public Window and anti-closure baseline
Public Window Protocol: An Open Future and Three Constraints
The Public Window protects the conditions for choosing, negotiating, refusing, changing one’s mind, exiting and rolling back before risks become irreversible. Its three constraints are window protection, preservation of experience, and non-manipulative cooperation.
Translation scope
This is a translation of the existing public web page. A description of a work is not its full manuscript. Source publication and version labels describe the Chinese work; they do not claim an English full-text edition is available. Unavailable English texts and downloads are marked explicitly.