Replies: 4 comments 1 reply
-
|
Short 30min course from the EU about the concept of Law as Code for free: https://academy.europa.eu/courses/understanding-law-as-code/view/
|
Beta Was this translation helpful? Give feedback.
-
|
Beta Was this translation helpful? Give feedback.
-
|
Good morning everyone, |
Beta Was this translation helpful? Give feedback.
-
|
Repo of main Demo from the presentation: https://github.com/benjamin-arfa/oll_2026_law_as_code |
Beta Was this translation helpful? Give feedback.


Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
1, The Challenge: Bridging the "Translation Gap"
The evolution of the modern administrative state has reached a critical inflection point where the volume and complexity of legal norms frequently outpace the capacity for manual implementation within digital systems. In the Swiss context, characterized by a highly granular federalist structure and a direct democratic mandate that frequently updates social and fiscal parameters, the traditional "write then program" approach results in significant friction, implementation delays, and potential for interpretive error. The "Law as Code" or "Rules as Code" (RaC) paradigm proposes a fundamental shift: representing legislation not merely as natural language text but as a structured, machine-readable logic model that can be interpreted by both humans and digital systems. By leveraging the upcoming Open Legal Lab we have a unique opportunity to prototype an automated pipeline that utilizes Large Language Models (LLMs) and declarative programming frameworks to bridge the divide between statutory intent and executable code.
For example: Legal norms like the Federal Act on Direct Federal Tax (DFTA) and Social Security (AHV) are updated frequently by popular vote, yet their digital implementation in administrative systems remains a slow, manual, and error-prone process.
This challenge asks: Can we use Large Language Models (LLMs) to automatically transform natural language Swiss law into verifiable, machine-executable code?
We are moving from "linear text" to "executable logic," turning the law itself into a shared digital infrastructure for a modern, transparent state.
2. Objectives & Mission
Our goal for the OLL 2026 is to build a functional Proof-of-Concept (PoC) pipeline that:
3. The Technical Approach: Declarative AI & RAG
To solve the problem of "brittle" LLM outputs, we will pioneer a cutting-edge technical stack:
4. Impact: Why Join Us?
By succeeding, we provide a blueprint for "Digital-Ready Legislation." Imagine a future where Swiss laws are published simultaneously as human-readable text and as an official logic "kernel" that powers all administrative software.
We are looking for:
Resource Suite Provided: Fedlex SPARQL access, Catala/OpenFisca language specs, and a curated financial test dataset for validation.
Let’s code the law. Join us at OLL 2026 to build the unified reasoning infrastructure for the Swiss digital state.
Ressources:
Attachments:
260408_Swiss Law as Code through LLM_AHW.pptx
260408_Swiss Law as Code through LLM_short_AHW.pptx
Beta Was this translation helpful? Give feedback.
All reactions