European Parliament Disables AI Assistants on Official Devices Over Data Security Fears

The European Parliament has disabled the built-in artificial intelligence assistants on all official devices used by lawmakers and staff. The decision, communicated in an internal email obtained by Politico, stems from unresolved concerns about where sensitive parliamentary data might end up.

The institution's IT service stated it cannot guarantee the security of information uploaded to the external servers of AI providers. The exact nature of what data is shared with companies like OpenAI, Microsoft, or Anthropic remains under review. "It is considered safer to keep such features disabled," the directive concluded.

This precaution highlights a core tension. When a user interacts with a cloud-based AI chatbot, the data can be subject to U.S. laws, potentially allowing authorities to request user information from the companies. Furthermore, the standard practice of using uploaded data to refine AI models raises the prospect of confidential details being inadvertently exposed.

The ban emerges amid broader European scrutiny of U.S. technology firms. These companies remain bound by American legislation, including recent aggressive use of administrative subpoenas by the U.S. Department of Homeland Security to obtain user data from platforms like Google and Meta. This legal environment has prompted several EU nations to reassess their reliance on U.S. digital infrastructure.

The Parliament's move stands in contrast to recent legislative proposals from the European Commission, which sought to ease data rules to facilitate AI training within the bloc—a plan criticized by some as a concession to American tech dominance. For now, within the Parliament's walls, human analysis will have to suffice.

Source: TechCrunch