---
title: "In public-sector AI, the contract decides access, and capability comes second."
url: https://wakenex.com/notes/in-public-sector-ai-the-contract-decides-access/
language: en
type: note
section: "Systems Research"
published: 2026-10-25T00:36:31+07:00
author: "WAKENEX"
id: "note:in-public-sector-ai-the-contract-decides-access"
x_post: https://x.com/AAwakened32198/status/2114048403164598413
alternate_language: https://wakenex.com/notes/in-public-sector-ai-the-contract-decides-access/index.zh.md
license: "© WAKENEX. You may quote and summarise with attribution and a link to the canonical URL."
cite_as: "WAKENEX. \"In public-sector AI, the contract decides access, and capability comes second.\" wakenex.com, 2026-10-25. https://wakenex.com/notes/in-public-sector-ai-the-contract-decides-access/"
---

# In public-sector AI, the contract decides access, and capability comes second.

**In one line:** In public-sector AI, the contract decides access, and capability comes second.

![A bright, sunlit government checkpoint under a clear sky. A plain, dull grey faceless figure is waved through an open turnstile glowing warm gold because it carries one oversized folder sealed with a big gold wax stamp; beside it, two much taller, ornate figures sculpted from glass and gold filigree — clearly more powerful — stand empty-handed behind a lowered barrier, shut out. White caption: Smarter, still banned. A Chinese line sits below it: 更聪明，却被挡在门外.](https://wakenex.com/assets/notes/in-public-sector-ai-the-contract-decides-access/cover-1600.jpg?v=6679fa1193)
*Image: WAKENEX*

Sweden proved it on October 24. Its Government Offices banned two consumer AI tools, ChatGPT and Claude, from staff computers and kept just one, Microsoft Copilot — and the block had nothing to do with which model is smarter. Copilot even runs Claude: Microsoft began offering Anthropic's models inside it this year, disabled by default in the EU. So the same weights a government bans can still reach it through an approved vendor.

What cleared Copilot was a contract and a security review Microsoft agreed to harden. ChatGPT and Claude were never put through that review. The Government Offices' director of digitalisation, Mikael Wettercrantz, said it plainly: using AI without a formal agreement has "consequences for our information security," because the agreement sets what the provider can and cannot do with a civil servant's inputs. A consumer chatbot ships that query to a server where a foreign government can later compel it; a contracted service names who is liable when it leaks.

So "best model" and "approved vendor" are two different games, and people keep conflating them. A lab can top every benchmark and still be shut out of an entire government, while its model quietly runs inside the tool that got in.

The honest counter: Copilot's tenant isolation and admin controls are genuine security engineering, and the review was substantive. Fair. But Sweden's stated basis was the missing agreement, and it never measured the others against it.

If you build on AI, the lesson is blunt: for regulated buyers your moat is the data-processing agreement and the jurisdiction you can run in. Capability gets you demos; contracts get you deployed.

This breaks if a government procurement openly picks the top-benchmark model over a contracted but weaker vendor.

---
Cite as: WAKENEX. "In public-sector AI, the contract decides access, and capability comes second." wakenex.com, 2026-10-25. https://wakenex.com/notes/in-public-sector-ai-the-contract-decides-access/

Web page: https://wakenex.com/notes/in-public-sector-ai-the-contract-decides-access/