Swedish Government Blocks ChatGPT and Claude on Work Computers Over Security Concerns

Swedish Government Implements AI Ban on ChatGPT and Claude

Sweden's Government Offices has blocked staff from using OpenAI's ChatGPT and Anthropic's Claude on work computers, a decision that took effect today.1

1

The Swedish government's move addresses mounting security concerns about how sensitive government data is handled when civil servants use unregulated consumer AI services. Staff can no longer reach the two chatbots through web browsers, with the organization steering them toward Microsoft Copilot instead.1

1

Mikael Wettercrantz, the Government Offices' director of digitalization and Chief Information Officer, informed staff about the AI ban through the internal network. He emphasized that AI tool usage without proper enterprise contracts creates significant data safety risks for Sweden's information security.2

2

The decision follows a review of AI tools used by workers at the Swedish Transport Administration earlier this year, which revealed "problematic" uses of AI platforms for government work.3

3

Data Privacy Concerns Drive Policy Shift

The core issue centers on how consumer AI services handle information entered by users. When government employees input confidential details into these platforms without formal agreements, that data can be fed back to developers and stored for training purposes.2

2

This creates serious GDPR compliance issues, particularly when data moves beyond European digital borders to be processed in data centers located in the US.2

2

Wettercrantz stated: "We have seen several examples of problematic use of open AI services. The examples concern both the sensitivity of the information that has been entered and how the AI-generated responses have been handled."4

4

US law can also allow authorities to demand data from companies under certain circumstances, adding another layer of concern for Swedish officials managing sensitive government data.4

4

Microsoft Copilot Emerges as Approved Alternative

While blocking ChatGPT and Claude, the Swedish government has designated Microsoft Copilot as the approved AI tool for official work. The Government Offices reviewed Copilot and strengthened its security settings, establishing a formal security agreement with Microsoft that addresses data protection concerns.3

3

This enterprise contract provides the safeguards and audit capabilities that were missing with the consumer AI services.2

2

However, questions remain about the technology underpinning this approved solution. Microsoft added Anthropic's Claude models to Copilot in early 2026, though these models are switched off by default in the European Union and require administrators to enable them.4

4

The Government Offices has not confirmed whether Claude models are enabled in its version of Copilot, creating some uncertainty about the distinction between blocked and approved services.4

4

Limited Access Through Swedish Platform

Some staff can still access ChatGPT and Claude through Intric, a Stockholm-based company that runs a platform for building AI assistants. Intric secured a contract with the Government Offices in July, positioning its platform as a secure alternative that runs on servers in Sweden without relying on foreign cloud providers.1

1

This approach allows limited AI governance while maintaining data privacy within Swedish borders, though officials have not disclosed how many employees can use this platform or what restrictions apply.4

4

Broader Pattern Across Swedish Agencies

The Swedish Tax Agency and the National Board of Institutional Care have already implemented similar restrictions on consumer AI tools.3

3

The Swedish Enforcement Authority is currently reviewing whether to adopt comparable measures.1

1

An internal audit at the Transport Administration found that around 10 AI services were being used in a widespread and risky manner, highlighting the scale of shadow IT challenges facing government agencies.4

4

This isn't Sweden's first move to restrict AI platforms. The country previously limited use of DeepSeek, the Chinese AI chatbot, over data security fears.1

1

Last week, the US went further and banned DeepSeek products across its federal civilian agencies.1

1

The European Parliament adopted an even broader approach in February, disabling AI features across its work devices, including ChatGPT, Claude and Copilot, because it could not guarantee the security of data handled by these services.4

4

OpenAI and Anthropic have not publicly commented on the Swedish decision.1

1

Watch for how other European governments respond to similar data privacy challenges as AI tool usage continues expanding across public sector organizations.