Sweden's Government Offices Block ChatGPT and Claude, Keep Microsoft Copilot

Sweden's Government Offices, which supports the prime minister and cabinet, will block staff from using ChatGPT and Claude on work computers starting Monday. Government workers will be told to use Microsoft's Copilot as their only approved AI tool, according to Sveriges Radio, Sweden's national public radio broadcaster, which receives public funding.

Mikael Wettercrantz, the Government Offices' director of digitalisation, explained the reason in a post on the staff intranet. He said that using AI services without an official contract could have "consequences for our information security."

"We have seen several examples of problematic use of open AI services. The examples concern both the sensitivity of the information that has been entered and how the AI-generated responses have been handled," he said.

Consumer AI tools without official agreements send users' questions to servers run by outside companies. U.S. law also allows authorities to require those companies to hand over data when asked.

At a minimum, an official security agreement explains what a provider can and cannot do with information entered by government employees. This is why a reviewed and approved business contract matters when choosing an AI tool, rather than simply using a public service.

Copilot is allowed because the Government Offices reviewed it and increased its security settings after the review. It did not do the same for ChatGPT or Claude. Some employees can still use both blocked tools through an AI platform built inside the Government Offices. However, the press office has not said how many employees can access them or what limits apply.

Sweden's Tax Agency and the National Board of Institutional Care have already introduced similar bans on consumer AI tools. The Enforcement Authority, which collects debts and makes sure court orders are carried out, is preparing its own decision. A separate internal review at the Transport Administration found that about 10 AI services were being used there in a "widespread and risky way."

In February, the European Parliament went further and turned off AI features on all its work devices, including ChatGPT, Claude, and Copilot. It said it could not guarantee that the data handled by these tools would remain secure.

AI Fears Grow as Germany, France, Sweden, and Canada Voice Concerns

European government concerns about AI data security have grown since Italy banned ChatGPT in 2023. Sweden is now taking a more limited approach than the European Parliament in Brussels: it is allowing one U.S. technology company's product while blocking two others.

Microsoft added Claude models to Copilot in early 2026. In the European Union, where Sweden is a member, Anthropic's models are turned off by default and must be deliberately enabled by system administrators.

Sweden's Government Offices has not confirmed whether Claude is enabled in its version of Copilot. This means the difference between the AI tool that is "banned" and the one that is "approved" may not be as clear as the announcement suggests.