AI Governance
AI Field Guide
Compare the institutions behind rival proposals: who can inspect, who can restrict access, and who has authority to set the rules. These are editorial comparisons, not inferred relationships between your Foundation and AI results.
§1
What this module measures, and what it does not
The Compass is a structured reader for AI governance instincts. It is not a technical exam, expertise credential, or full map of AI politics.
What it measures
Governance instincts under pressure
Whether your default answer is caution, capability, coordination, public oversight, state capacity, or openness when the AI case becomes concrete.
Tradeoffs across eight axes
Risk horizon, deployment pace, oversight, geopolitics, openness, military role, legitimacy, and human future.
How policy logics cluster
The archetype is the closest modeled family of answers, not a claim that the field has six natural camps.
What it does not fully cover
Technical expertise
The Compass does not certify whether someone understands model training, eval design, security engineering, or deployment operations.
Institutional feasibility
It asks what kinds of governance you find persuasive; it does not prove that those institutions can be built on the needed timeline.
The whole politics of AI
Several important perspectives are present only at the edge of the model. They belong in the reading shelf, not hidden in fine print.
§2
Governance dimensions
The vocabulary the AI module uses. Names, one-line description, and opposing poles only. It assigns no positions or scores.
Risk horizon
Do you prioritize current harms first, frontier catastrophic risk first, or insist on holding both in view?
Present harms firstFrontier risks firstDeployment pace
Do you learn through release and iteration, rely on threshold guardrails, or favor stronger precaution before deployment?
Iterate through useSlow at thresholdsPublic oversight
Should labs mostly govern themselves, or should states and public institutions supervise more directly?
Lab-led governancePublic authorityCompetition vs coordination
Do you see rivalry as the binding constraint, or coordination as the only durable answer?
Coordination firstCompetition firstOpenness vs control
Should advanced systems diffuse widely, or should access remain controlled and staged?
Broader accessControlled accessMilitary role
How bounded or how extensive should defense and military uses of frontier AI be?
Civilian restraintBounded defense useLegitimacy and rule-setting
Who should decide: technical insiders, governments, publics, or multilateral institutions?
Technocratic discretionPublic legitimacyHuman future
Do you treat transformative AI futures as promising, dangerous, constrained, or morally ambiguous?
Transformative futuresHuman control
§3
Where nearby archetypes part ways
Pairs whose fingerprints overlap on most axes. The split on one or two axes is what tells them apart.
§4
How IR and AI cross-read
Three meeting points between the IR Foundation and the AI Compass. They are related layers, not parallel scoring systems.
IR concept
Identity and legitimacy
In IR, the live question is whether legitimacy is procedural (rule-following) or substantive (outcome-defending), and across whose institutions it has to hold.
AI · Legitimacy and rule-setting: On the AI side, the same question becomes: who has standing to audit, license, or pause a model, and what grants that authority.
IR concept
Security rivalry
In IR, competition can be self-fulfilling. A security-dilemma reading turns defensive moves into offensive ones, and escalation follows.
AI · Competition vs coordination: On the AI side, that lens asks how much pace pressure is real strategic constraint and how much is invoked to dismiss safety arguments.
IR concept
Markets and dependence
In IR, the political-economy lens reads supply-chain coercion, market dependence, and who sets the terms of access.
AI · Openness vs control: On the AI side, the same lens reads compute, weights, and data concentration as governance questions, not only market ones.
§5
Perspectives the current model only partly covers
These are not hidden scored archetypes. They are adjacent shelves to keep in view when the six modeled families feel too narrow.
labor / automation
Labor, automation, and bargaining power
AI governance is also workplace governance: who captures productivity gains, who bears displacement, and whether workers have real leverage over deployment.
Would your AI result change if worker voice, not frontier safety or state capacity, were the starting point?
surveillance / platform / power
Surveillance and platform power
AI systems can deepen monitoring, prediction, and behavioral control by firms and states long before frontier-risk questions appear.
Where does your result treat concentration as a safety problem, and where should it treat concentration as a political-power problem?
environment / compute
Environmental and compute externalities
Compute, data centers, water, energy, minerals, and grid demand are material constraints, not background conditions.
Would your threshold for deployment change if compute externalities were visible on every capability decision?
global / south / sovereignty
Global South dependence and sovereignty
Many states will meet AI through imported cloud, foreign model access, compliance costs, and infrastructure dependence rather than frontier-lab choice.
Does your result assume the actor has the capacity to govern, or ask what happens when it does not?
chinese / official / framings
Chinese official framings
Chinese governance language combines sovereignty, development, safety, state capacity, and multilateralism in ways that do not map cleanly onto U.S. debate categories.
Which parts of your result are about AI governance itself, and which are about the political vocabulary you are used to reading?
open / source / anti / incumbency
Open-source and anti-incumbency politics
Open access can be a technical practice, a market strategy, a civil-liberties claim, or a challenge to state-lab concentration.
When you defend openness, are you defending innovation, accountability, sovereignty, competition, or all of them at once?
§6
Where to start reading
Three small entry points. Pick the one that matches why you arrived.
From the Compass
Read your nearest neighbor
Your archetype card already names a neighbor. The useful comparison is where two archetypes share most of the fingerprint but split on one axis.
Open the AI Atlas →From IR
Start with the cross-read
Three meeting points where IR vocabulary and AI governance vocabulary line up. The translation is uneven on purpose.
Open the cross-read →Browsing
Use the compact map
The Atlas maps governance patterns. The Field Guide explains each one without adding another assessment.
Open the AI Atlas →Three reading prompts used inside Field Guide entries to test how each archetype responds to a different shape of the next decade. They are not predictions and they do not enter any score on any page in this product.
Browse the whole Futures catalogue or consider your preferred conditions. This separate exercise leaves your AI Governance result unchanged.
Concentration tightens
Open weights hold
Multilateral baseline holds