@article{Simbeck, author = {Simbeck, Katharina}, title = {They shall be fair, transparent, and robust: auditing learning analytics systems}, series = {AI and Ethics}, volume = {4}, journal = {AI and Ethics}, number = {2}, publisher = {Springer Nature}, issn = {2730-5953}, doi = {10.1007/s43681-023-00292-7}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-19076}, pages = {555 -- 571}, abstract = {In the near future, systems, that use Artificial Intelligence (AI) methods, such as machine learning, are required to be certified or audited for fairness if used in ethically sensitive fields such as education. One example of those upcoming regulatory initiatives is the European Artificial Intelligence Act. Interconnected with fairness are the notions of system transparency (i.e. how understandable is the system) and system robustness (i.e. will similar inputs lead to similar results). Ensuring fairness, transparency, and robustness requires looking at data, models, system processes, and the use of systems as the ethical implications arise at the intersection between those. The potential societal consequences are domain specific, it is, therefore, necessary to discuss specifically for Learning Analytics (LA) what fairness, transparency, and robustness mean and how they can be certified. Approaches to certifying and auditing fairness in LA include assessing datasets, machine learning models, and the end-to-end LA process for fairness, transparency, and robustness. Based on Slade and Prinsloo's six principals for ethical LA, relevant audit approaches will be deduced. Auditing AI applications in LA is a complex process that requires technical capabilities and needs to consider the perspectives of all stakeholders. This paper proposes a comprehensive framework for auditing AI applications in LA systems from the perspective of learners' autonomy, provides insights into different auditing methodologies, and emphasizes the importance of reflection and dialogue among providers, buyers, and users of these systems to ensure their ethical and responsible use.}, subject = {Fairness}, language = {en} } @techreport{SimbeckWehnerFernseletal., author = {Simbeck, Katharina and Wehner, Marius and Fernsel, Linda and Schmodde, Lynn and Kalff, Yannick and Georgy, Jasmin and Resner, Monika}, title = {Leitfaden zur {\"U}berpr{\"u}fung von KI im Bildungsbereich auf Fairness}, organization = {HTW Berlin, Heinrich Heine Universit{\"a}t D{\"u}sseldorf}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-18115}, pages = {16}, abstract = {Der Auditierungsprozess von KI-Systemen im Bildungsbereich umfasst die Abgrenzung der Systembestandteile, die Identifizierung m{\"o}glicher Risiken, die Definition von Auditierungskriterien, die Vorbereitung der {\"U}berpr{\"u}fung sowie die {\"U}berpr{\"u}fung der Auditierungskriterien und ein kontinuierliches Monitoring. Die Auditierungskriterien sind dom{\"a}nenspezifisch und angelehnt an die sechs Prinzipien f{\"u}r ethische Learning Analytics Anwendungen nach Slade und Prinsloo (2013). Zur {\"U}berpr{\"u}fung der Kriterien k{\"o}nnen unter anderem die Dokumentation des KI-Systems gesichtet, der Quellcode analysiert, die dem Algorithmus zugrundeliegenden Daten untersucht und datenbasierte Tests durchgef{\"u}hrt werden. Die Ziele des Audits und der erw{\"u}nschte Nutzen sollten an die Lernenden kommuniziert werden. Nicht alle Lernenden haben den gleichen Wissensstand in Bezug auf KI-Systeme und k{\"o}nnen Bedenken gegen{\"u}ber dem Einsatz von diesen haben (Ifenthaler \& Schumacher, 2019). Die Pr{\"u}fung des Systems durch eine externe Institution kann einen positiven Einfluss auf die Wahrnehmung der Lernenden haben. Lernende empfinden ein System als fairer, wenn sie nachvollziehen k{\"o}nnen, dass das KI-System gepr{\"u}ft wurde. Lernende nehmen KI-Systeme umso positiver wahr, je mehr Informationen sie erhalten - auch wenn sie auf m{\"o}gliche Nachteile hingewiesen werden. Die transparente Offenlegung m{\"o}glicher Nachteile (z. B. technische Grenzen) beim Einsatz eines KI-Systems in der Bildung ist aus ethischer Sicht unabdingbar. Durch regelm{\"a}ßige Audits wird sichergestellt, dass die KI-Systeme in technischer Hinsicht funktionieren und gleichzeitig fair sind. Zugleich wird die Akzeptanz durch die Studierenden verbessert. Universit{\"a}ten profitieren daher von Audits ihrer KI-Systeme.}, subject = {Einf{\"u}hrung}, language = {de} } @article{KoechlingRiazyWehneretal., author = {K{\"o}chling, Alina and Riazy, Shirin and Wehner, Marius Claus and Simbeck, Katharina}, title = {Highly Accurate, But Still Discriminatory}, series = {Business \& Information Systems Engineering}, volume = {63}, journal = {Business \& Information Systems Engineering}, number = {1}, publisher = {Springer Fachmedien Wiesbaden}, issn = {2363-7005}, doi = {10.1007/s12599-020-00673-w}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-17115}, pages = {39 -- 54}, abstract = {The study aims to identify whether algorithmic decision making leads to unfair (i.e., unequal) treatment of certain protected groups in the recruitment context. Firms increasingly implement algorithmic decision making to save costs and increase efficiency. Moreover, algorithmic decision making is considered to be fairer than human decisions due to social prejudices. Recent publications, however, imply that the fairness of algorithmic decision making is not necessarily given. Therefore, to investigate this further, highly accurate algorithms were used to analyze a pre-existing data set of 10,000 video clips of individuals in self-presentation settings. The analysis shows that the under-representation concerning gender and ethnicity in the training data set leads to an unpredictable overestimation and/or underestimation of the likelihood of inviting representatives of these groups to a job interview. Furthermore, algorithms replicate the existing inequalities in the data set. Firms have to be careful when implementing algorithmic video analysis during recruitment as biases occur if the underlying training data set is unbalanced.}, subject = {K{\"u}nstliche Intelligenz}, language = {en} } @article{Simbeck, author = {Simbeck, Katharina}, title = {Publisher Correction: They shall be fair, transparent, and robust: auditing learning analytics systems}, series = {AI and Ethics}, volume = {4}, journal = {AI and Ethics}, number = {2}, publisher = {Springer Nature}, address = {Cham}, issn = {2730-5953}, doi = {10.1007/s43681-023-00301-9}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-19855}, pages = {573 -- 573}, language = {en} } @inproceedings{FernselKalffSimbeck, author = {Fernsel, Linda and Kalff, Yannick and Simbeck, Katharina}, title = {Audits for Trust: An Auditability Framework for AI-Based Learning Analytics Systems}, series = {Proceedings of the 17th International Conference on Computer Supported Education - Volume 2: CSEDU}, volume = {2}, booktitle = {Proceedings of the 17th International Conference on Computer Supported Education - Volume 2: CSEDU}, publisher = {SciTePress}, doi = {10.5220/0013254300003932}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-20790}, pages = {51 -- 62}, abstract = {Audits contribute to the trustworthiness of Learning Analytics (LA) systems that integrate Artificial Intelligence (AI) and may be legally required in the future. We argue that the efficacy of an audit depends on the auditability of the audited system. Therefore, systems need to be designed with auditability in mind. We present a framework for assessing the auditability of AI-integrating systems in education that consists of three parts: (1) verifiable claims about the validity, utility and ethics of the system, (2) evidence on subjects (data, models, or the system) in different types (documentation, raw sources and logs) to back or refute claims, (3) means to validate evidence such as technical APIs, monitoring tools, or explainable AI principles must be accessible to auditors. We apply the framework to assess the auditability of the Learning Management System Moodle, which supports an AIintegrating dropout prediction system. Moodle's auditability is limited by incomplete documentation, insufficient monitoring capabilities, and a lack of available test data.}, language = {en} } @inproceedings{FernselKalffSimbeck, author = {Fernsel, Linda and Kalff, Yannick and Simbeck, Katharina}, title = {Where Is the Evidence? A Plugin for Auditing Moodle's Learning Analytics}, series = {Proceedings of the 16th International Conference on Computer Supported Education (CSEDU 2024)}, volume = {2}, booktitle = {Proceedings of the 16th International Conference on Computer Supported Education (CSEDU 2024)}, publisher = {SciTePress}, doi = {10.5220/0012689800003693}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-20818}, pages = {262 -- 269}, abstract = {The paper presents the work-in-progress development of a Moodle plugin to improve the auditability of Moodle's Learning Analytics component. Future legislation, such as the EU AI Act, will require audits and "conformity assessments" of "high-risk" AI systems. Educational applications can be considered high-risk systems due to their important role in individual life and career paths. Therefore, their correctness, fairness, and efficiency must be assessed. However, auditing of the Learning Analytics functions in Moodle is limited. No suitable test-data is available, models and configurations are not persistent and only aggregated quality metrics are returned that are insufficient to assess fairness. The plugin addresses these issues and provides a data interface to extract data for audits. The plugin allows to a) upload and select data for the audit, b) clearly differentiate between model configuration and trained models, c) keep trained models, their configuration and underlying data for future inspections and comparisons, and finally, d) the plugin saves raw predictions for further analysis. The plugin enables the audit of Moodle's Learning Analytics and its underlying AI models and contributes to increased fairness and trustworthiness of Learning Analytics as well as its legally compliant application.}, language = {en} } @unpublished{SimbeckSchauerFernsel, author = {Simbeck, Katharina and Schauer, Sophie and Fernsel, Linda}, title = {The System Admin's Perspective: A Discussion on AI in Education with LMS Admins}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-20139}, pages = {10}, abstract = {Many higher education institutions (HEIs) are implementing or reviewing the implementation of predictive Learning Analytics to improve learning processes, identify students at risk or provide personalized learning paths. The responsibility for selecting, reviewing, implementing and supporting those systems falls on system administrators, an under-researched stakeholder group of Artificial Intelligence in higher education. In this paper, we summarize qualitative insights from a workshop with system administrators in German HEIs. We find that the system administrators are highly aware not only of system requirements but also of the needs of various shareholder groups such as institutional leadership, learners and educators and that they put high emphasis on ethical, transparent and compliant system use. We conclude that system administrators should be involved more in research on the use of technology in education and that AI systems used in education need to provide possibilities to sufficiently test the system, including anonymous yet realistic test scenarios and data.}, subject = {Moodle}, language = {en} } @inproceedings{KalffSimbeck, author = {Kalff, Yannick and Simbeck, Katharina}, title = {Explained, yet misunderstood: How AI Literacy shapes HR Managers' interpretation of User Interfaces in Recruiting Recommender Systems}, series = {Proceedings of the 5th Workshop on Recommender Systems for Human Resources (RecSys in HR 2025)}, booktitle = {Proceedings of the 5th Workshop on Recommender Systems for Human Resources (RecSys in HR 2025)}, editor = {Kaya, Mesut and Bogers, Toine and Bied, Guillaume and Johnson, Chris and Decorte, Jens-Joris}, publisher = {CEUR}, issn = {1613-0073}, url = {http://nbn-resolving.de/urn:nbn:de:0074-4046-X}, pages = {1 -- 10}, abstract = {AI-based recommender systems increasingly influence recruitment decisions. Thus, transparency and responsible adoption in Human Resource Management (HRM) are critical. This study examines how HR managers' AI literacy influences their subjective perception and objective understanding of explainable AI (XAI) elements in recruiting recommender dashboards. In an online experiment, 410 German-based HR managers compared baseline dashboards to versions enriched with three XAI styles: important features, counterfactuals, and model criteria. Our results show that the dashboards used in practice do not explain AI results and even keep AI elements opaque. However, while adding XAI features improves subjective perceptions of helpfulness and trust among users with moderate or high AI literacy, it does not increase their objective understanding. It may even reduce accurate understanding, especially with complex explanations. Only overlays of important features significantly aided the interpretations of high-literacy users. Our findings highlight that the benefits of XAI in recruitment depend on users' AI literacy, emphasizing the need for tailored explanation strategies and targeted literacy training in HRM to ensure fair, transparent, and effective adoption of AI.}, subject = {Explainable AI}, language = {en} } @inproceedings{SchauerSimbeckKuehn, author = {Schauer, Sophie and Simbeck, Katharina and K{\"u}hn, Patricia}, title = {Generative AI in Museums: Design Concepts for an Accessible Digital Mediation Station for a Renaissance Altar}, series = {Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering}, booktitle = {Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering}, publisher = {Springer}, address = {Heidelberg}, doi = {https://doi.org/10.1007/978-3-031-97254-6_20}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-22703}, pages = {283 -- 295}, abstract = {Exploring the potential of artificial intelligence (AI) in museum settings, this study finds concepts for a digital and inclusive mediation station giving context to a Renaissance sandstone altar at the Staatliches Museum Schwerin. Today, many museum visitors have less theological knowledge and come from diverse, often non-Christian, backgrounds. Thus, understanding and appreciating historical artworks and artefacts becomes more difficult. Therefore, exploring ways to present cultural heritage in a manner that is understandable, playful, and inclusive, regardless of the audience's origin or beliefs, is necessary. Through a collaborative week-long workshop involving communication design students, conceptual frameworks and prototype designs were iteratively developed and evaluated, laying the groundwork for further refinement and subsequent integration into the museum's permanent exhibition.}, language = {en} } @inproceedings{SchauerKatharinaFernsel, author = {Schauer, Sophie and Katharina, Simbeck and Fernsel, Linda}, title = {The System Admin's Perspective: A Discussion on AI in Education with LMS Admins}, series = {International Conference on Information Technology Based Higher Education and Training (ITHET)}, booktitle = {International Conference on Information Technology Based Higher Education and Training (ITHET)}, publisher = {IEEE}, doi = {https://doi.org/10.1109/ITHET61869.2024.10837664}, url = {http://nbn-resolving.de/urn:nbn:de:kobv:523-22759}, pages = {1 -- 5}, abstract = {Many higher education institutions (HEIs) are implementing or reviewing the implementation of predictive Learning Analytics to improve learning processes, identify students at risk or provide personalized learning paths. The responsibility for selecting, reviewing, implementing and supporting those systems falls on system administrators, an under-researched stakeholder group of Artificial Intelligence in higher education. In this paper, we summarize qualitative insights from a workshop with system administrators in German HEIs. We find that the system administrators are highly aware not only of system requirements but also of the needs of various shareholder groups such as institutional leadership, learners and educators and that they put high emphasis on ethical, transparent and compliant system use. We conclude that system administrators should be involvedmore in research on the use of technology in education and that AI systems used in education need to provide possibilities to sufficiently test the system, including anonymous yet realistic test scenarios and data.}, language = {en} }