European Parliament Restricts AI Tools on Official Devices Over Data Security Concerns

17.02.2026

The European Parliament has implemented a ban on AI-powered tools integrated into lawmakers' official devices, citing significant cybersecurity and privacy risks associated with uploading confidential correspondence to cloud-based services.

According to internal communications from the Parliament's IT department, the organization cannot guarantee the security of data transmitted to AI service providers' servers. The department stated that the full scope of information shared with AI companies is "still being assessed." Consequently, the directive concluded that "it is considered safer to keep such features disabled."

The decision addresses concerns that data uploaded to AI platforms—including Anthropic's Claude, Microsoft's Copilot, and OpenAI's ChatGPT—could be subject to U.S. legal jurisdiction, enabling American authorities to compel these companies to disclose user information. Additionally, AI chatbot platforms typically leverage user-submitted data to enhance their models, creating potential exposure scenarios where sensitive information uploaded by one user could inadvertently be accessed by others.

Europe maintains some of the world's most stringent data protection regulations. However, the European Commission recently proposed legislative amendments aimed at relaxing certain data protection requirements to facilitate AI model training on European citizens' data. This proposal has drawn criticism from privacy advocates who argue it represents capitulation to major U.S. technology corporations.

This restriction on AI tool access coincides with several EU member states reevaluating their dependencies on U.S. technology providers, which remain subject to American legal frameworks and the current administration's policies. Recent developments include the U.S. Department of Homeland Security issuing hundreds of subpoenas to technology and social media companies, demanding user information about individuals critical of administration policies.

Several major platforms, including Google, Meta, and Reddit, have reportedly complied with these requests despite the subpoenas not being judicially issued or court-enforced.

Sources:
Politico - EU Parliament blocks AI features over cyber privacy fears
Reuters - EU to ease AI privacy rules

Tags: data privacy AI security cybersecurity AI Regulation European Parliament

Share: VK Telegram Twitter