Vol-4263
urn:nbn:de:0074-4263-x

Copyright © 2026 for the individual papers by the papers' authors. Copyright © 2026 for the volume as a collection by its editors. This volume and its papers are published under the Creative Commons License Attribution 4.0 International (CC BY 4.0).

ExTraSafe 2026
Explainability, Transparency, and Safety 2026

Proceedings of the 1st Workshop on Explainability, Transparency, and Safety
co-located with 49th German Conference on Artificial Intelligence (KI2026)

Bremen, Germany, August 11, 2026.

Edited by

Benjamin Paaßen ¹
Vera Schmitt ²
Thomas Kosch³
Julius Schöning⁴
Hanna Drimalla¹
Florian Rabe⁵
Tim Schrills⁶
Daniel Neider⁷
Peter Fettke⁸

¹ Bielefeld University, Faculty of Technology, 33619 Bielefeld, Germany
² Technical University of Berlin, Department of Electrical Engineering and Computer Science, 10587 Berlin, Germany
³ Humboldt-University of Berlin, Department of Computer Science, 12489 Berlin, Germany
⁴ Osnabrück University of Applied Sciences, Faculty of Engineering and Computer Science, 49076 Osnabrück, Germany
⁵ University Erlangen Nuremberg, Technical Faculty, 91058 Erlangen, Germany
⁶ University of Lübeck, Institute of Human-Centered Interactive Systems, 23562 Lübeck, Germany
⁷ TU Dortmund University, Department of Computer Science, 44227 Dortmund, Germany
⁸ German Research Center for Artificial Intelligence (DFKI) and Saarland University, Institute for Information Systems, 66123 Saarbrücken, Germany

Table of Contents

Preface
Summary: There were 14 papers submitted for peer-review to this workshop. Out of these, 6 papers were accepted for this volume.

Session 1: Compliance, ethical aspects, and regulation

Technical, Organizational and Ethical Aspects of Integrating Enterprise Large Language Models into Existing Information Systems Infrastructure 1-8
Lara Kuhlmann, Alexander van der Staay, Maximilian Nebel, Annika Zemke, Christian Janiesch, Michael Henke
From Plausible to Verifiable: Hybrid Data as a Reference Standard for Explainable and Compliant AI 9-18
Julius Schöning, Paul Wachter

Session 2: Robustness and Trust

Robustness Verification of Traffic Sign Recognition: Practical Feasibility from a Safety Perspective 19-31
Marvin Schneider, Jörn Schneider, Holger Voos, Jürgen Kaiser
LLM Inconsistency under Paraphrase Attacks 32-48
Jiaao Li, Matthis Großkreutz, Tobias Peters, Ingrid Scharlau, Benjamin Paaßen

Session 3: Explainability

A New Paradigm for the Empirical Evaluation of the Understandability of LIME Visualizations 49-58
Marc Halbrügge
The Conversational Core Job as a Local Context in Game Explanations 59-70
Vivien Lohmer, Josephine B. Fisher

2026-09-08: submitted by Benjamin Paaßen, metadata incl. bibliographic data published under Creative Commons CC0
2026-09-14: published on CEUR Workshop Proceedings (CEUR-WS.org, ISSN 1613-0073) |valid HTML5|