Sweden’s Government Offices will block access to OpenAI’s ChatGPT and Anthropic’s Claude on official work computers from October 26, 2026, directing civil servants toward approved artificial intelligence tools, including Microsoft Copilot.
According to Swedish public broadcaster Sveriges Radio, the decision is intended to ensure employees use AI services that have been reviewed and approved by the administration. Officials have raised concerns about the information security risks associated with using AI tools without formal agreements.
The restriction reflects growing concerns about the sensitive information employees might enter into publicly available AI services and how that information could be handled. Government officials have warned that using services without appropriate agreements could create security risks.
Microsoft Copilot will remain an approved option after the Government Offices reviewed the service and strengthened its security settings. However, the administration has not publicly detailed all the changes made during the review.
The decision does not amount to a complete ban on ChatGPT and Claude across Sweden. Some government employees will retain access to both models through an internally developed AI assistant platform operated by the Government Offices. The exact number of employees who will retain access and the conditions governing that access have not been disclosed.
Other Swedish public agencies have also been reviewing their use of artificial intelligence. The Swedish Tax Agency and the National Board of Institutional Care have introduced restrictions on AI tools, while the Enforcement Authority has been considering its approach.
An internal audit at the Swedish Transport Administration also identified widespread use of several AI services and raised concerns about employees potentially entering sensitive information into publicly available tools. The Government Offices had previously restricted the use of DeepSeek as well.
The move highlights how government institutions are increasingly distinguishing between consumer AI services and tools that have been approved for official work. Rather than abandoning AI entirely, Sweden is seeking to limit potential information security risks while allowing employees to use selected services.
The restriction is scheduled to take effect on October 26, 2026, and applies to Sweden’s Government Offices rather than the entire country. It does not constitute a nationwide ban on ChatGPT or Claude for ordinary users.
If this article helped you, please consider supporting our work. Every small contribution keeps Abijita.com independent and running.
Sweden Blocks ChatGPT and Claude on Government Computers Over Security Concerns
