The Swedish government is blocking ChatGPT and Claude on staff computers starting today, Monday. Civil servants will now have to use Microsoft Copilot as their only approved AI tool, according to Sveriges Radio. The decision aims to protect sensitive government information from security risks linked to the use of AI services. Mikael Wettercrantz, the Government Offices’ director of digitalisation, informed staff about the move through the internal network. He warned that using AI tools without proper contracts could create information security risks. The government has reviewed Copilot and adjusted its security settings for the use.
Wettercrantz said officials had identified several cases of employees using AI services in ways that could put information at risk. “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 was quoted as saying in the report.
Consumer AI tools may send users’ questions and other information to servers operated by their providers. This can create concerns when government employees enter confidential details. US law can also allow authorities to demand data from companies under certain circumstances. Formal agreements help define how AI providers can use information entered by government staff.
The Government Offices reviewed Microsoft Copilot and strengthened its security settings. It did not carry out the same approval process for ChatGPT and Claude. However, some employees may still access the two blocked services through an internal AI platform developed by the Government Offices. Officials have not explained how many employees can use this platform or what restrictions apply.
Meanwhile, the decision raises questions about the technology behind the approved AI tool. Microsoft added Anthropic’s Claude models to Copilot in early 2026. In the European Union, these models are switched off by default and require administrators to enable them.
The Government Offices have not confirmed whether Claude models are enabled in its version of Copilot. This leaves some uncertainty about the difference between the AI services Sweden has blocked and the one it has approved.
Sweden’s Tax Agency and the National Board of Institutional Care have already introduced similar restrictions on consumer AI tools. The Enforcement Authority is also preparing its own decision.
Separately, an internal audit at the Transport Administration found that around 10 AI services were being used in a widespread and risky manner.
The European Parliament adopted a broader approach in February. It disabled AI features across its work devices, including ChatGPT, Claude and Copilot, because it could not guarantee the security of the data handled by these services.