Sweden: Government Favors Copilot, Blocks ChatGPT and Claude

The Swedish government offices cut access to ChatGPT and Claude on their staff computers from Monday, October 26. Only Microsoft Copilot remains authorized, after a security review. Except that Copilot can run Claude models.

Data Security, Why Stockholm Tightens the Screws

This decision is part of a broader trend. Several administrations seek to regulate the use of artificial intelligence, as illustrated by DeepSeek’s withdrawal from US federal civilian agencies. Italy had also blocked ChatGPT in 2023.

In Sweden, public radio revealed the measure, confirmed internally. Mikael Wettercrantz, director of digitization of government offices, explained on the intranet that using AI services without formal contracts could compromise information security.

The services notably noted problematic uses related to the sensitivity of data transmitted and the processing of the generated responses.

The issue thus concerns both the terms of use and the models themselves. With a public service, agents send their requests to an external provider without always controlling data processing. A contract allows, on the other hand, to define guarantees and responsibilities.

This is precisely what distinguishes Copilot from the two banned tools. After examining Microsoft’s solution, the administration strengthened its security settings before authorizing it. Some agents can nevertheless access ChatGPT and Claude via an internal AI platform. The press office did not specify how many people benefit nor under what conditions.

Other Swedish Administrations Follow the Movement

The restriction does not concern only government offices. The Tax Agency and the Institutional Care Council have already adopted comparable measures. The Enforcement Authority, responsible notably for debt recovery, is preparing its own decision.

Moreover, an internal audit of the transportation administration identified about ten AI services used widely and presenting risks. These findings illustrate the gap between rapid adoption of these tools and the implementation of adapted rules.

Other European institutions are also moving in this direction. In February, the European Parliament disabled AI functions on its professional devices, including those of ChatGPT, Claude, and Copilot, due to insufficient data protection guarantees. In the United States, Anthropic lost its appeal against the Pentagon in September, amid tensions around the use of its technologies.

Copilot Authorized, Claude Potentially Present

The distinction made by Stockholm deserves closer examination, however. Microsoft integrated Claude models into Copilot at the beginning of 2026. In the European Union, their activation requires a voluntary administrator intervention as they are disabled by default.

However, the Swedish government offices have not indicated if their version of Copilot allows using these models. The banning of Claude and authorizing Copilot could therefore coexist without the technical boundary being as clear as it seems.

In reality, Stockholm seems to favor tools whose usage conditions the administration controls rather than banning a technology altogether. It remains to be seen what precise guarantees justified the choice of Microsoft.

DISCLAIMER

The views, thoughts, and opinions expressed in this article belong solely to the author, and should not be taken as investment advice. Do your own research before taking any investment decisions.