Artificial Intelligence in the Service of Parliaments: Some Arguments for the Necessity of a Normative Legal Framework at the European Level

This contribution examines the need for European normative legal framework governing artificial intelligence implementation in parliamentary institutions. Using a three-pillar methodology—analysing soft law instruments (particularly Inter-Parliamentary Union Resolution and Guidelines), investigating European parliamentary AI use cases, and developing policy recommendations—the study reveals significant regulatory gaps. Whilst the EU AI Act provides limited coverage of parliamentary AI applications, several European legislative bodies (Estonia, Italy, Portugal, Spain, the European Parliament) have successfully implemented AI for speech recognition, document processing, and citizen engagement. Current soft law frameworks establish important principles including transparency, privacy protection, and sustainability, yet lack binding force. The research demonstrates that parliamentary AI's connection to democratic processes and national sovereignty necessitates stronger oversight. Two regulatory approaches are proposed: reclassifying parliamentary AI as high-risk under the EU AI Act, or developing dedicated European legislation addressing fundamental rights protection, national sovereignty, security concerns, and innovation promotion. The study concludes that establishing explicit European normative regulation would provide essential democratic safeguards whilst maintaining national sovereignty, parliamentary autonomy and fostering technological innovation in European democratic institutions.