Sweden’s Government Offices, which includes the Prime Minister’s Office and government ministries, has blocked staff from using OpenAI’s ChatGPT and Anthropic’s Claude on their work computers. The restriction took effect today.
Ekot, the news service of Swedish public broadcaster Sveriges Radio, first reported the plan on 23 October. The Government Offices later confirmed to SVT that staff would no longer be able to access the two chatbots through web browsers.
Instead, employees are being directed to Microsoft Copilot. The Government Offices said it had reviewed the tool and taken steps to make it safer for official use.
The move aims to ensure that staff use only AI tools approved by the government. Mikael Wettercrantz, head of digital services at the Government Offices, told Ekot that using AI tools without a formal agreement could put government data at risk.
However, some employees can still access ChatGPT and Claude through Intric, a Stockholm-based company that provides a platform for building AI assistants. Intric secured a contract with the Government Offices in July. The company says its platform runs on servers in Sweden and does not require foreign cloud providers.
Other Swedish government agencies have introduced similar restrictions. According to Ekot, the Swedish Tax Agency and the National Board of Institutional Care have also blocked certain AI services. SVT reported that the Swedish Enforcement Authority is considering taking similar action.
Sweden had already restricted the use of DeepSeek, a Chinese AI chatbot, before introducing the latest measures. Last week, the United States also banned DeepSeek products across its federal civilian agencies, according to BleepingComputer.
Large organisations in other countries are also tightening controls on external AI tools. In the United States, financial services company Charles Schwab has blocked AI agents, including Meta’s Muse, from carrying out actions such as moving client money.
These steps reflect growing concerns among governments and businesses about data security and the risks of allowing AI tools to access sensitive information. Many organisations are now choosing to limit access to tools that have not been formally approved or assessed for workplace use.
