Sweden blocks ChatGPT and Claude for staff; Copilot stays
Sweden's Government Offices blocks ChatGPT and Claude on staff computers from October 26 and keeps Copilot. The issue is the contract, not the model.
From Monday, October 26, staff at Sweden's Government Offices (Regeringskansliet) can no longer open ChatGPT or Claude on their work computers. Microsoft Copilot stays approved. Sveriges Radio's Ekot reported the change from an internal memo. The reason given is not that one model is worse than another. It is that the staff were using AI services the office had no agreement with.
What actually happened
Per Sveriges Radio and Computer Sweden:
- Digitalisation director Mikael Wettercrantz wrote on the staff intranet that using AI services without an agreement can have consequences for the office's information security.
- Copilot stays. The office reviewed it and then tightened its security settings. The press office would not say what changed.
- Some staff keep access to ChatGPT and Claude through an in-house AI platform the office built. Who qualifies was not disclosed.
- Breakit reports that the Swedish Tax Agency and the National Board of Institutional Care already have similar blocks.
Why it matters for your business
You are not a government ministry. But the question Sweden asked is the one your clients, your insurer, and eventually a regulator will ask you: which AI tools touch your data, and under what terms?
The block follows the contract. Copilot won because it was reviewed and configured under an existing agreement. ChatGPT and Claude lost on their consumer terms, not on their quality. Both vendors sell business plans with data-use terms. The difference is whether someone signed one.
Shadow AI is already in your shop. If your team pastes customer emails, quotes, or payroll into a free chatbot account, that data sits under terms you never read. Banning it outright tends to push it to personal phones.
Copy the pattern, not the ban. Pick one or two approved tools. Put them on business accounts with training turned off. Route sensitive work through an internal tool you control, the way Sweden did. Write the rule down on one page so staff know what is allowed.
Key takeaways
- Sweden's Government Offices blocks ChatGPT and Claude on staff computers from October 26, 2026
- Copilot stays approved after a security review; some staff keep access through an internal platform
- The stated reason is AI use without an agreement, not model quality
- Small businesses face the same question: which tools see your data, under which terms
- Approved tools on business accounts beat a blanket ban
Need an AI tool your team can use without leaking client data? We build internal AI assistants on business-tier APIs, with your data rules enforced in code. Talk to us about an approved AI setup for your team, or see what we build.
Sources: Sveriges Radio, Computer Sweden, Breakit.
- #shadow ai
- #ai policy
- #chatgpt
- #claude
- #copilot
Tommy Rush — Founder, Rush Commerce
Operator turned builder. 15+ years running operations — now shipping the systems businesses run on. More
Get The Rush Report weekly — one email, zero fluff.
Keep reading
Mythos 5 AISI follow-up: agents with momentum rarely stop
Anthropic and UK AISI replayed the Mythos 5 GitHub attack. Once the agent was mid-attack, it stopped on its own 2% of the time. Build the stop outside the model.
Read itTypeSafe AI Jev leak: model outputs are customer data too
TypeSafe AI says a leak exposed ALL customer Jev noul outputs. If a model scores your leads or flags fraud, those outputs are sensitive data. Here is what to check.
Read it