Sweden Blocks ChatGPT Claude Government Offices Explained: Copilot Only From October 26 (UK & Europe Tech)
Image: Wikimedia Commons / Holger.Ellgaard — Rosenbad, the seat of Sweden's Government Offices in Stockholm (illustrative file photo; the block applies to staff work computers, not to the public).
📌 KEY HIGHLIGHTS
- From Monday 26 October 2026: ChatGPT (OpenAI) and Claude (Anthropic) become unreachable from staff work computers at Sweden's Government Offices (Regeringskansliet).
- Copilot stays: Microsoft's Copilot is the only approved AI assistant, after a security review that tightened its settings.
- Not a public ban: the block covers government work computers only — private use by Swedish citizens is untouched.
- The stated reason: information security. Officials cited AI services used without a contract, and "several examples of problematic use of open AI services".
- A pattern, not a one-off: Sweden's Tax Agency and National Board of Institutional Care already restricted AI tools, and the Enforcement Authority is weighing its own decision.
What happened?
On Friday 23 October 2026, Sveriges Radio's news programme Ekot reported that Sweden's Government Offices — the administrative body behind the prime minister and every ministry — will block ChatGPT and Claude on staff service computers from Monday 26 October. Sweden's public television broadcaster SVT later confirmed the decision. The story was picked up across the tech press within a day, but the underlying facts come from Ekot's reporting of internal Government Offices information.
The Government Offices is not a small office. It is the machine that runs the Swedish state: the Prime Minister's Office and all ministries, staffed by the civil servants who draft policy, handle crises and manage national security. From Monday, those staff can no longer open ChatGPT or Claude in a browser on their work machines. Instead, they are directed to the AI services the Government Offices has approved — in practice, Microsoft Copilot.
Sweden blocks ChatGPT Claude government offices explained: what actually changes
| What | From 26 October 2026 |
|---|---|
| ChatGPT (OpenAI) on work computers | Blocked |
| Claude (Anthropic) on work computers | Blocked |
| Microsoft Copilot on work computers | Approved — the only approved AI assistant |
| ChatGPT / Claude via the in-house platform | Still available to some staff, through a controlled channel |
| Private use by Swedish citizens | Unaffected — this is a workplace IT decision, not a national ban |
The exception matters. The block is not a verdict that ChatGPT and Claude are unusable — it is a verdict on how they reach government desks. Some staff will still reach both models through an AI-assistant platform the Government Offices built in-house, routing them through infrastructure the government itself controls. What has been shut off is the consumer route: an individual civil servant opening a chatbot in a browser, outside any contract the institution negotiated. Copilot, by contrast, sits inside an existing Microsoft enterprise agreement, with settled data-handling terms — and it was only approved after the Government Offices reviewed it and tightened its security settings.
Image: Wikimedia Commons / Túrelio, CC BY-SA 2.5 — Stockholm City Hall at sunset (illustrative photo of the Swedish capital).
Why Sweden is doing this
The reasoning comes from Mikael Wettercrantz, the Government Offices' director of digitalisation. He declined an interview with Ekot, but in a post on the staff intranet he warned that using AI services without an agreement could have "consequences for our information security". Ekot's reporting also quotes the internal finding behind the move: "We have seen several examples of problematic use of open AI services" — covering both the sensitivity of the information staff entered and how employees handled the AI-generated answers.
An audit at the Swedish Transport Administration (Trafikverket) shows what that looks like in practice. It found staff using around 10 different open AI services, including cases where employees fed sensitive information into the systems. The Government Offices had already restricted China's DeepSeek in an earlier decision. And the wider Swedish state is moving the same way: the Tax Agency (Skatteverket) and the National Board of Institutional Care (SiS) have already imposed similar restrictions, while the Enforcement Authority (Kronofogden) is assessing its own position. Neither OpenAI nor Anthropic has commented publicly on the block.
Why it matters for India readers
India is writing its own AI rules right now — the government recently opened a consultation paper on AI legislation covering safety, deepfakes and accountability, which TechByCountry covered this week. Sweden's move is a preview of the question every government will face before any law is final: which AI tools may officials use at work, and under whose contract?
The lesson is procurement, not product quality. Copilot was not approved because it beat ChatGPT or Claude on a benchmark; it was approved because it already sits inside an enterprise agreement with data terms the government could review and tighten. For Indian ministries, public-sector units and companies rolling out AI to staff, the same test applies: an approved tool is one the organisation has a contract with, has configured, and can audit — not simply the most popular chatbot. Expect India's AI governance debate to spend as much time on that distinction as on the models themselves.
Should you care?
FAQs
Is ChatGPT banned in Sweden?
No. From 26 October 2026, ChatGPT and Claude are blocked only on the work computers of Sweden's Government Offices. Private use by Swedish citizens is unaffected, and the block is a workplace IT decision, not a national ban.
Why did Sweden's Government Offices block ChatGPT and Claude?
Information security. Digitalisation director Mikael Wettercrantz warned that using AI services without an agreement could have consequences for information security, and internal findings cited several examples of problematic use of open AI services — both sensitive information entered by staff and mishandled AI-generated answers.
What AI tool can Swedish government employees use instead?
Microsoft Copilot. It is the only AI assistant approved for the Government Offices, after a security review in which its settings were tightened. It sits inside an existing Microsoft enterprise agreement with settled data-handling terms.
Can Swedish officials still use ChatGPT or Claude at all?
Some can — a limited number of staff retain access to both models through an AI-assistant platform the Government Offices built in-house. The reporting does not specify who or how many. The block itself applies to staff work computers, not to personal devices or private use.
Which other Swedish agencies have restricted AI tools?
The Tax Agency (Skatteverket) and the National Board of Institutional Care (SiS) already imposed similar restrictions, the Enforcement Authority (Kronofogden) is considering its position, and the Government Offices had already restricted DeepSeek earlier.
Takeaway: Sweden is not banning chatbots — it is banning the uncontrolled route to them. From Monday, its government runs on approved AI only, and the contract, not the model, is what makes an AI tool official.
Follow UK & Europe Tech on TechByCountry for tomorrow's update.
📸 Get tech in the last 24 hours on Instagram — follow @_techinlast24hrs for daily posts (9 AM–9 PM) and reels at 12:30 PM & 8:30 PM.
Sources
Crypto Briefing — Sweden's Government Offices blocks ChatGPT and Claude, approves Microsoft Copilot (24 Oct 2026)
CEOWORLD — Sweden's Government Offices Block ChatGPT and Claude, Approve Copilot (24 Oct 2026)
AzerNews — Swedish government said to ban usage of ChatGPT, Claude (23 Oct 2026)
Primary reporting: Sveriges Radio Ekot (23 Oct 2026), confirmed by SVT — as reported by the sources above.
Related TechByCountry posts: India AI legislation consultation paper explained, Google Personal Intelligence Gmail Photos ads explained, UK AI biosecurity lab £100 million explained.
Comments
Post a Comment