Skip to content
Cisent Message us on WhatsAppMessage us

Blog & news

Threat researchCisent

Sweden's government blocked ChatGPT and Claude for its staff: there was no agreement to protect its data

Staff at the Swedish Government Offices can no longer use ChatGPT or Claude in the browser on their work computers; Copilot stays, because it does have a security agreement with the government.

Since Monday, October 26, 2026, the Government Offices of Sweden (Regeringskansliet) have blocked OpenAI’s ChatGPT and Anthropic’s Claude on their staff’s computers.1 The block covers the versions used in the browser.2 The government wants its staff to use only approved services.3

No agreement, no access

Mikael Wettercrantz, the Government Offices’ director of digitalization, wrote on the intranet that using AI services without an agreement can have “consequences for our information security.”1 According to Euronews, the concern is that sensitive information typed into a service without a data protection agreement could end up stored and used to train the model.2

The assistant that stays is Microsoft Copilot: it passed a security review and is covered by a security agreement between Microsoft and the Swedish government,2 which also tightened its security.3 Some staff will keep using ChatGPT and Claude, but only through an AI assistant platform the government built itself.3

Not just the government

According to Ekot, Swedish public radio’s news service, the Swedish Tax Agency (Skatteverket) and the National Board of Institutional Care (Statens institutionsstyrelse) have also blocked them, and the Swedish Enforcement Authority (Kronofogden) may follow.1

Why it matters

The decision is not against AI: the government kept an assistant under agreement and built a platform of its own. What it cuts off is the use of services nobody contracted. When someone pastes a document into an AI their organization never contracted, the organization doesn’t know where that information ends up or what it’s used for.

Where Cisent comes in

The underlying problem is not knowing what happens to the information typed into someone else’s AI.

Many AI security tools process your logs in third-party models. Cisent uses its own models, with no outside AI provider in between: the Manthys models, four families trained by Cisent. Mind reasons, Sentinel defends, Nova brings the scale and Reflex filters, and they work inside Cisent alongside Husker, the agentic layer. We don’t train on your data. Not even anonymized.

See how we protect your data

Sources

  1. Viktor Eriksson, Computer Sweden, “Regeringskansliet spärrar Chat GPT och Claude,” October 23, 2026, reporting Ekot, Sveriges Radio. https://computersweden.se/article/4239687/regeringskansliet-sparrar-chat-gpt-och-claude.html↩︎

  2. Jonathan Benton, Euronews, “Swedish government blocks ChatGPT and Claude. Here’s why,” October 26, 2026. https://www.euronews.com/2026/10/26/swedish-government-blocks-chatgpt-and-claude-heres-why↩︎

  3. Finwire in Börsvärlden, “Regeringskansliet spärrar användning av ChatGPT och Claude – Ekot,” October 23, 2026. https://borsvarlden.com/artiklar/regeringskansliet-sparrar-anvandning-av-chatgpt-och-claude-ekot↩︎