Sweden's Government Offices block ChatGPT and Claude on staff computers from Monday

Sweden’s Government Offices, Regeringskansliet, will block ChatGPT and Claude on staff work computers from Monday, 26 October, Sveriges Radio’s Ekot reported on 23 October. Staff are meant to use the AI services the Government Offices has approved, including Microsoft Copilot.

What Ekot reports

• The block. From Monday, ChatGPT and Claude can no longer be reached from staff work computers.
• The reason. Chief digital officer Mikael Wettercrantz declined an interview, but wrote on the Government Offices intranet that using AI services without an agreement can have “consequences for our information security” (Ekot’s quote, our translation).
• Copilot gets tighter, too. Internal information seen by Ekot says the Government Offices raised security in Copilot after a review. The press office would not describe the change, citing security.
• Not a total cutoff. The press office told Ekot that some staff will keep access to a separate platform for AI assistants, built by the Government Offices itself, through which ChatGPT and Claude can still be used.
• DeepSeek was already restricted there before this change.

SVT and Dagens industri carried the same details, both citing Ekot.

Other agencies

Ekot’s survey found the Government Offices is not alone. The Swedish Tax Agency (Skatteverket) and the National Board of Institutional Care (SiS) have already introduced blocks, and the Enforcement Authority (Kronofogden) is preparing a decision.

The Transport Administration (Trafikverket) is the cautionary example. Its internal audit found widespread, risky use of about ten different AI services earlier this year, and wrote that it had seen “several examples of problematic use of open AI services”, covering both how sensitive the input was and how the AI answers were handled (our translation).

What it does not say

Nobody has said whether the block covers phones or personal devices, which staff qualify for the internal platform, or what contract terms separate Copilot from the blocked apps. Neither OpenAI nor Anthropic is quoted in the reports.

Why it matters

The pattern is the one many companies already follow: shut off consumer chat apps, keep one contracted assistant, and give a smaller group a governed route to other models. What stands out is that a central government is doing it openly, with several agencies moving the same way. If you sell AI tools to the Swedish public sector, the contract and data terms matter more than the model.