In an internal email circulated on 16 February 2026, the European Parliament mandated the deactivation of built‑in AI features on work tablets and smartphones issued to lawmakers and staff, citing cybersecurity and data‑protection risks linked to cloud‑based data transfers.
The Parliament’s e‑MEP tech support desk issued a warning that some of the affected AI tools automatically send information to external servers even for tasks that could be processed locally on the device.
The email stated that, as these features continue to evolve and spread across more devices, the full scope of data shared with third‑party service providers remains unclear, making it “safer” to keep them turned off until a comprehensive assessment is completed. Affected functions include writing assistants, text‑summarization tools, enhanced virtual assistants, and webpage‑summary features. Core applications such as email, calendars, and standard document tools will remain available.
Officials emphasized that the move responds to fears that sensitive legislative and constituent data could be routed through foreign cloud infrastructure, potentially exposing it to jurisdictions where national laws — such as the US CLOUD Act — allow authorities to compel data access from American‑based providers. Under the EU’s General Data Protection Regulation (GDPR), transferring personal data outside the bloc requires strict safeguards, and Parliament staff reportedly do not believe current AI tools meet those standards.
The decision also reflects a broader push within the EU for digital sovereignty and reduced dependence on non‑European technology platforms. In 2023, the Parliament had banned TikTok on staff devices over similar security concerns, and in late 2025 a cross‑party group of MEPs had urged the institution to shift away from Microsoft software toward European‑developed alternatives.
Additionally, EU lawmakers have been encouraged to apply comparable precautions on personal devices used for work, including reviewing AI‑related settings, restricting app permissions, and avoiding the use of AI features that scan or analyze internal documents and work emails.
The Parliament’s press office said it constantly monitors cybersecurity threats, and takes swift action to prevent them. However, it declined to detail specific measures, describing them as sensitive.
