FOSDEM 2026 - Open Source & EU Policy

FOSDEM 2026 - HowTheyVote.eu - how we make European Parliament roll-call votes more accessible

Study Guide: eTools and AI in the EU Legislative Process

This study guide is designed to provide a comprehensive review of the digital tools (eTools) and artificial intelligence (AI) applications currently utilized or proposed for regulatory simplification and consistency within the European Union and its Member States. The material is based on detailed reports concerning the modernization of the legislative cycle through advanced technology.

Part I: Short-Answer Quiz

Instructions: Answer the following ten questions in 2 to 3 sentences each, based on the provided source context.

1. What are the three simplified stages of the legislative cycle used to categorize eTools, and what does the "Ideation" stage specifically involve?

2. What is LEOS (Legislation Editing Open Software), and what is its primary technical objective?

3. How does the Estonian "HANS" system assist in parliamentary operations, and what is its reported accuracy?

4. Explain the purpose of the "LLaMandement" system developed by France’s Directorate General of Public Finances.

5. What is "Akoma Ntoso" (AKN), and why is it considered critical for the future of digital legislation?

6. How does the "RIA Platform" in Greece differ from tools that use generative AI to conduct analysis?

7. What is the "Ulysses Suite," and what unique function does its "U7" module perform?

8. Identify two constitutional risks associated with the implementation of AI-driven tools in parliamentary work.

9. Why are there currently fewer eTools available for the "Evaluation" stage of the legislative cycle compared to the "Ideation" or "Development" stages?

10. What is the "functionality-first" roadmap proposed as a policy option for the European Parliament?

• -------------------------------------------------------------------------------

Part II: Answer Key

1. The three simplified stages are Ideation, Development, and Evaluation. The Ideation stage corresponds to the preparation phase of policy-making, involving the identification of regulatory needs and gaps through data collection, stakeholder input, and predictive analytics to assess hypothetical impacts.

2. LEOS is an open-source software developed by the European Commission to support the collaborative drafting and management of legislative texts. Its primary objective is to ensure legal consistency and create machine-readable output by enforcing document structures based on the Akoma Ntoso (AKN) XML standard.

3. HANS is a speech-to-text transcription system used by the Estonian Riigikogu to record plenary and committee debates. It achieves approximately 95% accuracy (a 4-5% word error rate) and allows for the publication of official verbatim records within an hour of a session ending.

4. LLaMandement is a fine-tuned Large Language Model (LLM) designed to generate neutral, legally precise summaries of French legislative amendments. It was created to address administrative backlogs caused by the rising volume of amendments, allowing staff to prepare bench memoranda and inter-ministerial meetings more efficiently.

5. Akoma Ntoso (AKN) is a machine-readable XML standard designed specifically for the structural representation of legal and legislative documents. It is critical for ensuring interoperability across different institutions and jurisdictions, enabling seamless data exchange and future-proofing digital transformation efforts.

6. The RIA Platform is a procedural Government-to-Government service that manages the workflow and document lifecycle of Regulatory Impact Assessments. Unlike analytical AI tools, it focuses on standardizing the process of preparing, signing, and submitting reports to ensure completeness and quality control before they reach the Greek Parliament.

7. The Ulysses Suite is an array of AI-driven modules developed by the Brazilian Chamber of Deputies to modernize legislative activities. Its "U7" module specifically handles authentication during deliberations by integrating facial recognition and biometric technology to verify the identity of members during virtual or hybrid voting.

8. Constitutional risks include the risk of undue influence by third parties, which endangers parliamentary autonomy, and the risk of technological malfunctioning, which could undermine the principle of continuity of power. Additionally, there are risks regarding technological disparity between parties and potential transfers of parliamentary powers to opaque AI systems.

9. The Evaluation stage lacks tools primarily due to a deficit of high-quality, granular data and a lack of a solid, common methodological foundation. Meaningful ex-post evaluation requires clear indicators and measurable KPIs (Key Performance Indicators) that are often missing or vaguely defined at the time of a law's adoption.

10. The functionality-first roadmap suggests prioritizing the development of eTools based on specific user-driven needs rather than adopting technology for its own sake. This approach focuses on high-impact functionalities like identifying regulatory gaps, ensuring internal consistency of texts, and simulating how amendments affect existing laws.

• -------------------------------------------------------------------------------

Part III: Essay Questions

Instructions: Use the provided source context to develop comprehensive responses to the following prompts. (Answers are not supplied).

1. The Digital Sovereignty Dilemma: Evaluate the implications of European parliaments utilizing non-European Large Language Models (LLMs) such as Meta’s LLaMA or OpenAI’s GPT. Discuss how this affects institutional autonomy and what steps are suggested to mitigate these risks.

2. Interoperability vs. Autonomy: Analyze the challenges of creating a "European Inter-Parliamentary Interoperability Framework." Discuss why Member States might resist common software solutions and how "shared building blocks" might offer a middle ground.

3. The Impact of Structured Data on "Better Regulation": Argue the case for why data standardisation (e.g., AKN4EU, EuroVoc) is a prerequisite for effective AI integration in law-making. Reference specific tools that rely on these standards for their functionality.

4. AI as a Collaborative Tool: Examine the concept of the "human-in-the-loop" and the "augmented LEOS" vision. How does the transition from "passive" tools to "intelligent co-authoring environments" change the role of legislative staff and MEPs?

5. Closing the Evaluation Gap: Propose a strategy for increasing the number and effectiveness of eTools in the Evaluation stage of the policy cycle. Incorporate lessons learned from the Slovenian MOPED system and the Greek RIA platform.

• -------------------------------------------------------------------------------

Part IV: Glossary of Key Terms