Greek law NLP pipeline reveals 'Illusion of Simplicity' with 45% vague terminology

A new research paper from Despoina Antonakaki and Sotiris Ioannidis introduces an NLP pipeline for semantic auditing of the Greek legislative process. The system combines parliamentary discourse analysis with automated scraping of Government Gazette documents using a headless browser to bypass anti-scraping barriers. Applying LLMs to 2025 Hellenic Parliament transcripts, the researchers extracted 534 unique law citations and used debate frequency as a proxy for political salience. Their analysis uncovered an 'Illusion of Simplicity'—laws presented as simplifications actually exhibit high structural complexity and ambiguity. Among 312 tagged ambiguity instances, 45% stem from vague terminology and 25% from deferred executive delegation.

The pipeline also introduces the Political Discrepancy Index (PDI), which measures alignment between ministerial promises and enacted law. Testing on three high-frequency laws (4808/2021, 4412/2016, 4662/2020) showed that the dominant outcome is 'Deferral'—commitments shifted to future Ministerial Decisions. Cross-reference network analysis confirmed a highly entangled legal system where foundational provisions are among the most frequently amended. The result is a semantically linked dataset and an interactive auditing interface that enables journalists, policymakers, and citizens to track legislative quality and accountability at scale.