By Hamza Faisal
From Monday, October 26, civil servants in Sweden's Government Offices will no longer be able to open ChatGPT or Claude on their work computers. The Regeringskansliet, the body that supports the prime minister and the country's ministries, is blocking both chatbots on staff machines and directing employees to approved tools instead, Swedish public broadcaster Sveriges Radio reported on Friday.
The decision makes Sweden one of the clearest cases yet of a European central government choosing a single approved AI supplier for everyday work — and switching off the two best-known alternatives on its own network. Microsoft's Copilot is the service that survives the cut: it is the only AI assistant to have passed the Government Offices' own security review, after which its security settings were tightened further.
A contract, not just a policy memo
The reasoning comes from Mikael Wettercrantz, the Government Offices' director of digitalisation. He declined an interview with Sveriges Radio's news programme Ekot, but warned on the staff intranet that using AI services without an agreement in place could have consequences for information security.
That distinction is the heart of the move. Staff opening ChatGPT or Claude on the web were using services with no contract governing how government information is handled. Copilot, by contrast, is a service the Government Offices has formally approved — and internal documents seen by Ekot show its security setup was upgraded after a review before it got the green light. The press office declined to detail those changes, citing security concerns.
The block is not a ban on AI at work, and it is not absolute. Some staff will keep reaching ChatGPT and Claude through an AI-assistant platform the Government Offices built itself, which routes the tools through a controlled environment. The press office did not say who qualifies or how many people that covers.
Part of a wider Swedish crackdown
The Government Offices are not acting alone. Sweden's Tax Agency, Skatteverket, and the National Board of Institutional Care have already introduced similar restrictions, and the Swedish Enforcement Authority is weighing the same step. Use of China's DeepSeek was restricted in an earlier decision.
The pattern extends beyond Stockholm's ministries. At the Swedish Transport Administration, an internal audit this year found staff using ten different open AI services, with auditors documenting cases where employees fed sensitive information into the systems and mishandled AI-generated answers. "We have seen several examples of problematic use of open AI services," the auditors wrote, pointing both to what staff entered and to how they treated the answers that came back.
A template other governments may copy
What makes Sweden's move worth watching beyond Scandinavia is that it pairs a network-level block with a sanctioned alternative that cleared a security review — the complete loop many organisations still have only half of. Ban the unsanctioned tools, then actually provide one that passed review, so employees have somewhere legitimate to go.
The contrast with the United States is instructive: Washington's order removing DeepSeek from federal civilian agencies targets a foreign model on supply-chain grounds, while Sweden's block applies to the two most widely adopted Western assistants on data-handling grounds. The criteria differ, but the destination is the same — government work consolidating onto whichever vendor can sign the security paperwork. On Monday morning in Sweden's ministries, chatgpt.com and claude.ai will stop responding on work machines. Copilot will not.
By Hamza Faisal
