Overlaps, contradictions and inconsistencies in rules create legal uncertainty, impose bureaucratic burden on businesses and hamper innovation—especially in highly regulated sectors. Artificial intelligence (AI) can reliably identify simplification potential of rules and support efforts to streamline the regulatory framework—provided the right conditions are in place. This is the conclusion of the “Feasibility study on an AI-based application for identifying the simplification potential of regulations,” conducted on behalf of the Federal Ministry of Finance (BMF) under the leadership of d-fine.
This interdisciplinary study combines research and practical perspectives from law-making and artificial intelligence. It demonstrates that such an AI-based application is feasible, provided the necessary technical and organisational foundations are in place:
1. Legal knowledge base: An AI application needs a comprehensive, up-to-date and quality-assured database of relevant rules and their interpretations to deliver reliable analyses. As a first step, the authors recommend focusing on publicly available legal norms.
2. Suitable methodological approach: The study identifies key requirements for successfully identifying and analyzing simplification potentials with AI and proposes suitable methodological approaches, each combining multiple AI techniques.
3. Reliable results: High reliability of the proposed approaches can be achieved by integrating automated preparation and validation steps. In addition, expert review of the AI results remains indispensable.
The study was produced under the direction of d-fine in cooperation with A&O Shearman, Fraunhofer IAIS, Lexemo and Prof. Dr. Florian Möslein (Philipps-University Marburg).
This interdisciplinary study combines research and practical perspectives from law-making and artificial intelligence. It demonstrates that such an AI-based application is feasible, provided the necessary technical and organisational foundations are in place:
1. Legal knowledge base: An AI application needs a comprehensive, up-to-date and quality-assured database of relevant rules and their interpretations to deliver reliable analyses. As a first step, the authors recommend focusing on publicly available legal norms.
2. Suitable methodological approach: The study identifies key requirements for successfully identifying and analyzing simplification potentials with AI and proposes suitable methodological approaches, each combining multiple AI techniques.
3. Reliable results: High reliability of the proposed approaches can be achieved by integrating automated preparation and validation steps. In addition, expert review of the AI results remains indispensable.
The study was produced under the direction of d-fine in cooperation with A&O Shearman, Fraunhofer IAIS, Lexemo and Prof. Dr. Florian Möslein (Philipps-University Marburg).
