Semantic annotation of parliamentary debates and legislative intelligence enhancing citizen experience

The concept of “Legislative Intelligence” (LegisIntel) refers to Artificial Intelligence (AI) and semantic analytics tools implemented in parliaments to enhance citizen experience in monitoring complex interrelations among various contents of parliamentary proceedings. The integration of a suite of digital solutions can build upon the core functionality of Semantic Annotation of Parliamentary Debates. Using well-established Natural Language Processing (NLP) technologies, linked to ontologies and Knowledge Graphs (KG), it can help identify the concepts and entities throughout texts, and index sentences and summaries as per a citizen’s customized knowledge base. These annotations can then be leveraged to recommend relevant text excerpts end-users could build upon, within teams if they chose to do so, and possibly compose and customize legislative critiques and recommendations thoroughly tested for coherence, accuracy, and evidence. The present study proposes an international open-source initiative among parliaments to ensure the launch and viability of a suite of LegisIntel solutions. It reports on the completed initial phase of this initiative, aiming to prepare discussions in launching an international consultation among peers. The goals of this phase are to document the core functionality of LegisIntel solutions and formulate a proposed architecture that may serve to generate ideas from various developer communities. The Action Design Research (ADR) methodology is used in this process, with results focused on system artefacts such as an interface mockup, a functional design, and a model of infrastructure components. The conclusion addresses risks and outlines the next steps of this initiative.