Public — HIME Global Public Capacity — English edition

Central thesis — provisioning opportunity, not automating administration

HIME Global Public Capacity is a concept for an international framework that begins in the world's public sector. Its aim is not to reduce public servants with AI, nor to pack more work into the same hours. It is to finish necessary public work in less time, at higher quality and lower burden — and to return the time thus created to learning, research, rest, culture, family, community, and international exchange. The axis of global competition shifts from working longer and processing more to finishing sooner, learning well, resting well, and living well.

Three layers — W100, Lex50, HIME

The concept has three layers. W100 (this site) is the coordinate system of knowledge: what each public servant knows, is learning, and can do. Lex50 is the coordinate system of law: which rights, duties, powers, procedures, exceptions, and remedies each task rests on. HIME is the execution platform that connects staff, AI, equipment, canonical records, authority, evidence, approvals, and outputs. Kept separate but connected, they distinguish what AI can do from what institutions permit.

The Time Dividend — returning the gains as time

Gains from AI must not be recovered solely through headcount cuts or added workload — this is the core norm. The allocation of created time is fixed before deployment: half is returned to staff themselves (shorter hours, holidays, leisure), a quarter goes to development, research, exchange, learning, and knowledge transmission, and the remaining quarter is at the organisation's discretion (service quality, surge capacity). Pay is not reduced when hours shorten. Public-sector trials of shorter hours (Iceland 2015–2019, the United Kingdom 2022, and others) show that time can be returned while quality is maintained.

Public technologists — if you cannot hire them, train them

Local governments and small public bodies around the world cannot hire AI engineers on the market. So the premise is reversed: train the people who already know the work. Through education inside working hours, every member of staff learns to decompose their own tasks into purpose, inputs, judgement, outputs, and exceptions; to separate what may be delegated to AI from what humans must adjudicate; to verify outputs; and to share workflows. Assessment is by attainment — what one can safely design, execute, and audit — recorded on W100 coordinates, independent of job titles, portable across borders.

Machines work at night; humans adjudicate in the morning

Around-the-clock operation exists so that machines work nights and holidays in order that humans do not. Work that ends in a draft — collection, classification, cross-checking, translation drafts — runs unattended; anything with legal effect (transmission, expenditure, licences, finalising canonical records) executes only after human approval. AI is never an authority holder. Each morning, staff find not a backlog but a docket of cases that AI has prepared and humans must decide.

Graduation — the public sector becomes a school

Public organisations are redesigned as institutions of education, research, and craft: people join, learn, leave results, and graduate into society. A well-trained member of staff moving on to the private sector, research, community work, or international cooperation is not a loss but public human-capital formation and technology transfer. Graduates carry their attainment records and authorship, and institutional paths exist for their return.

Inequity — shrinking the discount imposed by accident

People of equal ability face vastly different income, opportunity, and administrative quality by accident of country, region, language, and their organisation's equipment. Shrinking this unreasonableness is the final goal. Equipment, AI, education, and maintenance are provisioned as one internationally pooled service; participants choose not devices or models but a capacity class matched to their staff and workload. The goal is not uniform outcomes but guaranteed paths through which ability and rights are not unjustly lost to accident.

The one-person post stays inside the standard

Public service always contains posts that one person must run alone: the sole IT officer in a small municipality, a single posting at a field office, a lone specialist, a site cut off by disaster. Treating such posts as "too small to qualify" leaves the places that need support most outside the framework. Solo operation is therefore designed as a formal configuration, not an exception. The line is single: equipment may be reduced; structure may not. Shared compute, redundancy, around-the-clock operation — all may be omitted in proportion to scale. But which version of which canonical record a decision rested on and where, who approved it, that records cannot later be erased, and that work continues when AI stops: none of these may be omitted, however small the post. Being alone is no reason for a resident to lose the right to an explanation and to object.

One person cannot be both drafter and approver. Three measures compensate. Drafting and adjudication do not occur on the same day (tomorrow's self does not decide as yesterday's self did). A regional shared body performs periodic sampled review. And the AI roles are split between drafting and refutation. The third is an aid only — AI never approves in a human's place. Where a second pair of human eyes cannot be created inside, it is borrowed from outside. Solo operation is permitted; isolated operation is not.

At a one-person post, a successor is not always on hand. Continuity therefore rests in records rather than people: canonical records, workflows, evaluation criteria, and approval history live with the organisation, not on the officer's desk — so that a successor or relief officer can identify, on their first day, what must be taken over. Work existing only in one person's head and on one person's device is treated as an incident. Nor is this configuration a temporary state pending more staff. A one-person post is one person for structural reasons — population, the scarcity of a speciality, geography. Presume its dissolution, and that post is treated forever as a place not yet ready.

Phases

Phase	Content
Phase 0	Concept — shared terms, principles, authority, and the Time Dividend
Phase 1	Limited pilots — ten-person units in two to three countries
Phase 2	Routine work — documents, translation, procurement, accounting, statistics, statute checking
Phase 3	24-hour operation — night jobs adjudicated by humans each morning
Phase 4	Education and graduation — attainment records and circulation of people
Phase 5	International communities — daily co-production among same-occupation practitioners
Phase 6	Private-sector spillover — procurement conditions and compatible infrastructure

Status

This page is the public summary of the concept. The full design — institutions, technology, education, evaluation, and risks — is being prepared as a single design document that distinguishes facts, estimates, and proposals, and attaches grounds or assumptions to every number. The Japanese text is authoritative; this page is a reference edition, and further languages will follow ratification by participating organisations.