Sweden's government blocks ChatGPT and Claude on staff computers, keeps Copilot

Representative image: Pexels / Pixabay
Sweden's Government Offices blocked web versions of ChatGPT and Claude on staff computers from October 26, 2026, citing information security; Copilot stays.
What Happened
Sweden's Government Offices, which include the Prime Minister's Office and the government ministries, began blocking the web versions of ChatGPT and Claude on staff work computers on Monday, October 26, 2026, citing information security. The decision, first reported by Sveriges Radio and confirmed by the government's press office, steers civil servants to Microsoft Copilot, which has been through a security review and is covered by a security agreement with the government. Mikael Wettercrantz, the Government Offices' director of digitalisation, warned that using AI platforms without formal agreements could have consequences for information security.
Key Facts
- Who: the Swedish Government Offices, including the Prime Minister's Office and ministries.
- When: the block took effect on October 26, 2026.
- What is blocked: the public browser versions of OpenAI's ChatGPT and Anthropic's Claude on work computers. It is not a ban on the AI models for the public or for companies.
- Why: information-security concerns; reports cite "several instances of problematic use" of AI services that had no agreement behind them.
- Alternative: Microsoft Copilot, in an approved version that has been security-vetted and gives the government more control over settings.
- Exception: some staff can still reach the two models through an internal AI platform built by the Government Offices; how many, and under what limits, was not disclosed.
- Similar restrictions are already in place at the Swedish Tax Agency and the National Board of Institutional Care, with the Swedish Enforcement Authority reported to be next.
Why It Matters
Governments everywhere face the same problem. Generative AI tools such as ChatGPT and Claude can draft letters, summarise reports and translate documents in seconds, and civil servants have started using them, often through free public websites. But anything typed into a public chatbot leaves the government's own systems and is processed on a company's servers, under that company's terms. For a ministry handling draft laws, security matters or citizens' personal data, that is a serious risk.
Sweden's answer is not to stop AI use but to channel it. Officials say the issue is the lack of a formal agreement covering how data is handled, not the technology itself. Copilot was allowed because it has been checked and is covered by a contract that sets security terms and lets the government control settings. The internal platform that still gives some staff access to ChatGPT and Claude suggests the same principle: use the models through a controlled, contracted route rather than the open web.
The move also shows a wider trend in Europe. Public bodies are increasingly asking where their data goes and which company's rules apply to it.
For India, the debate is directly relevant. Indian ministries, state governments and public-sector companies are also adopting AI tools, and the Digital Personal Data Protection Act places duties on anyone handling personal data. Sweden's approach, a clear list of approved tools backed by security agreements and an internal platform for wider access, is one model Indian departments could study as they write their own rules.
Impact
Short-term: Swedish civil servants lose direct web access to ChatGPT and Claude at work and must use Copilot or the internal platform. More Swedish agencies may follow with similar restrictions.
Long-term: Expect more governments to move from informal AI use to approved, contracted tools, which favours AI providers that offer government-grade security terms and data controls.
Who is affected: Swedish government employees, AI companies competing for public-sector contracts, and officials in other countries, including India, who are drafting rules for safe AI use in government.
Key Takeaway
Sweden's Government Offices blocked the web versions of ChatGPT and Claude on staff computers from October 26, 2026, over information-security concerns, while keeping the security-vetted Microsoft Copilot.
Questions and Answers
Has Sweden banned ChatGPT and Claude?
No. Only the Swedish Government Offices have blocked the public web versions on staff work computers from October 26, 2026. The public and companies in Sweden can still use them.
Why did Sweden's government block ChatGPT and Claude?
Officials cited information security. Staff had been using AI services with no formal agreement on how data is handled, and the director of digitalisation warned this could have consequences for information security.
Which AI tool can Swedish government staff still use?
Microsoft Copilot, in an approved version that has been security-vetted and is covered by a security agreement. Some staff can also reach ChatGPT and Claude through an internal government AI platform.
Are other Swedish agencies doing the same?
Yes. The Swedish Tax Agency and the National Board of Institutional Care already have similar restrictions, and the Swedish Enforcement Authority is reported to be next.
Disclaimer: Prepared by the Peepals newsroom from publicly available sources with AI assistance. Information is accurate to the best of our knowledge at the time of publication and may change. Images may be representative. Not professional advice. Report an error via our contact page.
By Anika Sharma, AI & Innovations EditorCovers AI & Innovation and more for Peepals Tech. All stories by Anika Sharma
Brotli flaw behind Node2Shell spreads: Microsoft rushes .NET patches, Python affected
UP cabinet clears four GCC projects worth Rs 1,710 crore
Get the day's stories in your inbox
Peepals Weekly: the week decoded, free. One email a week, unsubscribe in one click.
Anika Sharma is a pen name of the Peepals Global Editorial Team, which researches, fact-checks and publishes every story. The portrait is an illustration.









