Research map / Large language models / Language, safety and interpretability
An Expert System for Danish Traffic Law Cases (Student Abstract)
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Presents FÆRDXEL, a Datalog-based expert system that uses SLD-resolution to answer and explain queries about Danish traffic law cases, with a UI for legal professionals.
Problem
Legal domains require explainable judgments, yet modern ML systems provide answers without traceable reasoning; expert systems for law have seen little recent research. The paper targets explainable answers to Danish traffic law questions.
Approach
A Datalog encoding of Danish traffic law (hand-derived from a commentary book, with reification and syntactic sugar compiled to CNF clauses) answered via SLD-resolution, whose refutation trace is converted into a natural-language legal explanation through an interface.
Key results
- Knowledge base encoded as 2579 clauses after CNF conversion of the Danish traffic law rules.
- In an early proof of concept used in a court case, FÆRDXEL reached the same conclusions as the sitting judge about which laws the defendant had broken, but could not give a punishment range.
- Feedback from a group of legal experts identified usability problems (Datalog input too difficult) and four potential end-user groups.
Limitations
- Punishment reasoning is not yet encoded; the tool could not provide a punishment range in the court case.
- Inputting facts requires Datalog, which legal experts found too difficult.
- No empirical soundness evaluation yet; only an early proof of concept.
Experimental setup
Datasets: Færdselsloven med Kommentarer (commentary book on Danish traffic law)
Builds on: SLD-resolution Datalog
Closest papers by meaning
- LexChain: Modeling Legal Reasoning Chains for Chinese Tort Case Analysis (AAAI 2026 · similarity 0.81)
- Interpretable Long-Form Legal Question Answering with Retrieval-Augmented Large Language Models (AAAI 2024 · similarity 0.81)
- LexEval: A Comprehensive Chinese Legal Benchmark for Evaluating Large Language Models (NeurIPS 2024 · similarity 0.80)
- Lawma: The Power of Specialization for Legal Annotation (ICLR 2025 · similarity 0.79)
- Freeing the Law with LOCUS: A Local Ordinance Corpus for the United States (NeurIPS 2026 · similarity 0.79)
- Argumentative Large Language Models for Explainable and Contestable Claim Verification (AAAI 2025 · similarity 0.79)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.