The European Parliament temporarily disabled built-in AI features on every work device issued to MEPs and staff. According to Politico, writing assistants, text summarisers, and virtual assistants were all switched off. This is a stark reminder of the mistrust and concern of AI systems embedded in government institutions.
What actually happened
The internal email, first reported by Politico, stated that some AI features use cloud services to carry out tasks that could be handled locally, sending data off the device. It added that the full extent of data shared with service providers is still being assessed, and until that is clarified, it is safer to keep such features disabled.
The disabled tools include writing assistants, summarisation tools, enhanced virtual assistants, and webpage summary features on tablets and phones. Core apps like email, calendar, and document editors continue to work normally. This is not a ban on AI in general. It is a targeted block on built-in features that transmit data to external cloud servers.
The Parliament also urged MEPs to apply similar caution on their personal devices, warning them to avoid feeding work emails or internal documents into AI tools that scan or analyse content.
Then Copilot got a bug
Just days after the Parliament’s decision, a separate story broke that underlined exactly why this kind of caution matters. Microsoft confirmed that a bug in its 365 Copilot Chat had been allowing the AI to read and summarise confidential emails since late January, bypassing the data loss prevention policies that organisations rely on to protect sensitive information.
The BBC spoke to Microsoft, which acknowledged the error. Here’s BBC reporter Shaun Dacosta.
According to BleepingComputer, the bug was first detected on 21 January. Copilot’s work tab was summarising emails stored in users’ Sent Items and Drafts folders, even when those emails carried confidentiality labels designed to block AI access. Microsoft says a fix has now been deployed worldwide for enterprise customers, but it took nearly a month from detection to resolution.
To be clear, the Parliament’s decision is not connected to this specific bug. The IT email went out before the Copilot story became public. But the two events together tell the same story: organisations do not yet have full visibility over what AI tools are doing with their data.
Why did the Parliament act
The European Parliament’s actions are a reminder of the bigger picture happening in the world of AI right now.
- Confidentiality: The European Parliament handles draft legislation, negotiation positions, and sensitive political correspondence daily. AI features that send that content to a US-hosted cloud infrastructure create an exposure that the institution cannot control.
- Jurisdiction. As TechCrunch reported, this decision lands at a moment when the US Department of Homeland Security has sent hundreds of subpoenas demanding US tech companies hand over user data. Google, Meta, and Reddit have reportedly complied in several cases. For European legislators, the idea that their working documents sit on servers subject to those demands is a serious concern.
- Sovereignty. This is part of a pattern of discussion now emerging across the EU. In 2023, the Parliament banned TikTok from staff devices over data protection fears. In November 2025, 38 MEPs signed a letter calling for the Parliament to replace Microsoft products entirely with European alternatives. The letter also mentions replacing non‑EU hardware, such as devices from Dell, HP, and LG, not just Microsoft software. Finnish MEP Aura Salla put it bluntly at the Open Source Policy Summit in Brussels earlier this month: the EU runs on Microsoft, and the US could turn us off inside one hour. According to a European Parliament report, the EU relies on non-EU countries for over 80% of its digital products, services, and infrastructure.
In the short term, staff lose tools that many offices were already quietly using for drafting, translation, and summarisation. More manual work, a step back to traditional workflows.
Then there is shadow AI. Research shows nearly 35% of what employees type into AI tools contains sensitive data. The Parliament can disable features on official devices, but it cannot stop someone from pasting a document into ChatGPT on their personal phone. The policy sets the boundary. Enforcing it is another matter.
We have a gap. There is no widely adopted EU-built AI assistant ready to drop in as a replacement. Mistral is mooted as one solution, and France is already using it in some government agencies.
Both of these stories point to the same conclusion. If you do not know where AI is sending your data, you have a governance problem.
If you want to build your AI literacy and skills, I can help.
📚 Grab AI Playbook on Amazon, the perfect podcast companion.
📰 Subscribe to Human in the Loop, my LinkedIn Newsletter
🎓 Take the CPD Accredited AI Playbook Diploma
📞 Book a call at digitaltraining.ie
Sources
Politico: “EU Parliament blocks AI features over cyber, privacy fears” (16 February 2026): politico.eu
TechCrunch: “European Parliament blocks AI on lawmakers’ devices, citing security risks” (17 February 2026): techcrunch.com
The Next Web: “The European Parliament pulls back AI from its own devices” (17 February 2026): thenextweb.com
The Register: “European Parliament bars lawmakers from AI tools” (17 February 2026): theregister.com
Key Statistics
- AI features disabled on all European Parliament work devices (tablets and smartphones) as of 16 February 2026
- Features affected: writing assistants, text summarisers, enhanced virtual assistants, webpage summary tools
- Core apps (email, calendar, documents) are unaffected
- TikTok was banned from Parliament devices in March 2023
- 38 MEPs signed a letter in November 2025 calling for the replacement of Microsoft with European alternatives
- EU relies on non-EU countries for over 80% of digital products, services, and infrastructure (European Parliament report)
- Amazon, Microsoft, and Google command 70% of the European cloud market
- 35% of employee inputs to AI tools contain sensitive data (Metomic research)
About the Podcast Host
AI and Communications Specialist | Author | Podcaster | International Speaker
Joanne Sweeney is a recognised leader in digital marketing and digital transformation, specialising in the practical application of AI to enhance communication strategies. The founder of Digital Training Institute, Joanne, empowers leaders and communications professionals to leverage AI responsibly and effectively, ensuring transparency, trust, and tangible results.
She completed the Harvard Kennedy School Executive Programme in AI and Public Policy this summer before launching her fourth book AI Playbook, in September. Joanne has also published two editions of Public Sector Marketing Pro, which were preceded by her first title, Social Media Under Investigation Law Enforcement and the Social Web.
Her books have become go-to resources for governments, public sector and policing agencies, as well as marketing and communications professionals, all navigating the challenges of the changing Internet.
Joanne is the host of the AI SIX Podcast, a daily six-minute podcast that provides actionable insights, expert interviews, and practical advice to help professionals navigate the new world of work.
