Sweden restricts government access to ChatGPT and Claude

Sweden’s Government Offices (Regeringskansliet) will block direct access to OpenAI’s ChatGPT and Anthropic’s Claude on employee work computers starting October 26, 2026. The decision aims to mitigate information security risks, specifically the potential for employees to enter confidential or work-related data into external AI platforms without appropriate contractual protections.

While direct access is being restricted, staff will not lose all AI capabilities. Some employees will continue to use these models through a dedicated AI assistant platform developed internally by the Government Offices. Additionally, Microsoft Copilot will remain a primary approved alternative, following a security review by the agency to strengthen its protections.

This move follows similar actions by other Swedish public authorities, including the Swedish Tax Agency (Skatteverket) and the National Board of Institutional Care (SiS). The Swedish Enforcement Authority is also reportedly preparing a similar decision.