All editions · Policymaker Edition

Your Community, Your AI — CC BY 4.0

AI Policy You Can Articulate

A Plain-Language Briefing for Policymakers and Elected Representatives

A five-part series for members of parliament, ministers, councillors, advisers, and officials who need to understand artificial intelligence well enough to articulate policy about it clearly — in the House, in committee, on a doorstep, or in a briefing to a colleague. It assumes no technical background. If you can read a select-committee paper, you can read these articles. (Any unfamiliar term in this series is defined in plain language in the glossary.)

This is an educational briefing, not a lobbying document, and it is deliberately cross-party. Where it draws on a specific policy proposal — In Our Own Hands, a draft published for public discussion — it treats that proposal as one worked example among the options, not a position to be adopted. That proposal is offered by its author as an independent contribution to public debate. It is not a party document and not a product, and no party, official, or individual has endorsed it. The purpose here is narrower and more useful to a legislator: to give you the underlying understanding, so that whatever position you take, you can state it precisely and defend it under questioning.

A note on scope. This edition is written for an international readership, and it uses Aotearoa New Zealand as its worked example — a small, sovereign, common-law democracy is a useful case precisely because the choices are stark. The fundamentals in Articles 1 to 4 hold in any jurisdiction. The specific instruments named — a privacy act here, a procurement rule there, a domestic AI charter — are illustrations: where you sit, the names change and the principles do not. Read each New Zealand reference as "and the equivalent in your own system."

The Series

1. What AI Actually Is (and What It Isn't)

The distinction that changes the policy problem: AI has moved from chatbots that answer to agents that act. What the engine does — predict the next word — why whether it "reasons" is genuinely unsettled, why you cannot discharge oversight by reading the machine's own account of itself, and why the questions that matter for legislation are whose patterns does it carry and whose hands are on the controls.

2. Big Tech AI and the Sovereignty Question

Why where an AI runs, and under whose law, is a question of state, not merely of procurement. The jurisdiction problem beneath the technology — the US CLOUD Act, FISA Section 702, the PRC National Intelligence Law — why data residency is not the same as data sovereignty, and why the argument is supplier-agnostic: not which foreign supplier to trust, but how to keep custody and control at home.

3. Why Principles Are Not Enough — The Governance Challenge

The load-bearing article. Why voluntary principles drift, why "training wears off," and why aspiration without architecture is insufficient. The responsibility gap and the moral crumple zone. Why the EU AI Act's human-oversight requirement should be read as structural, not promised — and why New Zealand's Algorithm Charter and Public Service AI Framework, being voluntary and non-binding, sit on the other side of that line. What Wittgenstein, Berlin, and Ostrom tell a legislator about the limits of rules.

4. What Community-Governed AI Actually Does Today

A factual inventory, so that policy is grounded in what the technology does rather than in hype or fear. What one community-governed system does in production today — grounded answers, bounded action, structural boundaries, confidence indicators — and, stated plainly, what is still under development.

5. The Policy Opportunity — A Menu, Not a Manifesto

The practical levers, as a menu any party could draw from: seven cross-party commitments, procurement built on jurisdiction of inference and data residency, a provenance-and-audit standard with a public register, a national capability fund, a reversibility requirement for publicly funded pilots, EU AI Act alignment for market access, deeper standards participation, and — on te Tiriti and Māori data — deference to Māori-led frameworks rather than a Crown prescription.

Who This Is For

These articles are written for people who make, scrutinise, or advise on public policy — MPs and their staff, ministers and officials, local-government members, party researchers, and committee advisers. No technical background is required. The aim is not to turn you into an AI specialist; it is to give you enough grounded understanding to form a view and articulate it clearly.

Regulatory Context

The series is written with reference to the frameworks a policymaker is most likely to be asked about:

• The EU AI Act (Regulation 2024/1689) — risk classification, transparency obligations, and human-oversight requirements; relevant to New Zealand as a condition of market access, not as an obligation.
• The General Data Protection Regulation (GDPR) — data protection, lawful basis, and the expectation of meaningful information about automated decisions.
• The Privacy Act 2020 (NZ) — the domestic baseline for personal-information handling.
• The Algorithm Charter for Aotearoa New Zealand (2020) and the Public Service AI Framework (2025) — voluntary, non-binding commitments; their voluntary status is a central theme of Article 3.
• ISO/IEC 42001 — the international standard for AI management systems, and the standards track (ISO/IEC JTC 1/SC 42) through which a small country can shape the global ground rules.

Want to use AI tools like these well, and safely? Our free courses — Working with Claude and Agents at Work — teach the practical skills, from getting trustworthy answers to deciding what to hand an agent. For the full technical architecture behind Village AI, see Village AI — Agentic Governance.