Navigating Generative AI in Education
Risks, Rewards, and the Role of Explainability

This symposium will explore how generative AI can enhance personalized learning experiences while tackling key challenges such as fairness, transparency, and bias in educational content. Featuring a distinguished panel of experts from leading institutions, this seminar promises to create valuable insights into the future of AI-driven education.

Interactive panel discussions will offer attendees the chance to engage directly with leading experts, encouraging dynamic and forward-thinking dialogue. This seminar is designed to be a platform for interdisciplinary collaboration, promoting a shared vision and research strategy that harnesses collective expertise to tackle the complex challenges posed by AI in education.

Key topics

• Personalized Learning and AI: Explore how machine learning models can customize educational experiences for individual learners, enhancing engagement and outcomes.
• Explainable AI (XAI): Understand the importance of transparency in AI algorithms to foster trust and ethical use in educational settings.
• Ethical and Legal Considerations: Investigate the intersection of AI with law and policy, addressing issues such as authorship, copyright, and the responsible adoption of AI tools in education.
• Digital Transformation in Teaching: Discover innovative approaches to integrating digital media and AI into curriculum development and teaching practices.

Program
January 21

09:30	Opening & welcome remarks Sebastian Bonhoeffer Collegium Helveticum
09:45	Digital Learning as a Dynamic Process Fostering Creativity and Curiosity Charlotte Axelsson Zurich University of Teacher Education (PHZH), CH
10:45	Coffee break
11:00	Modeling Human Behavior and Learning with Machine Learning Tanja Käser Swiss Federal Institute of Technology Lausanne (EPFL), CH
12:00	Biases in Generative AI for Education Francesco Sovrano Collegium Helveticum
13:00	Lunch break
14:15	AI in Computing Education Evaluating Effectiveness and Developing Support Tools Jaromir Šavelka Carnegie Mellon University, US
15:15	Coffee break
15:30	AI in Education, Research, and Innovation Scenarios and Recommendations Abraham Bernstein University of Zurich, CH
16:30	Legal and Didactical Implications of AI in Education Panel Discussion Moderated by Francesco Sovrano Margaritha Windisch ETH Zurich, CH Jaromir Šavelka Carnegie Mellon University, US Daniel Flück ETH Zurich, CH
17:30	Closing remarks Followed by a small reception

Program
January 22

09:30	Opening & welcome remarks Francesco Sovrano Collegium Helveticum
09:45	Learning with Digital Media: A Cognitive, Affective, and Social Process Sascha Schneider University of Zurich, CH
10:45	Coffee break
11:00	Visual Analytics Approaches for Exploring Large Language Models Rita Sevastjanova ETH Zurich, CH
12:00	From Concept to Classroom Insights into Designing Chatbots for Primary English Language Learning Luzia Sauer Zurich University of Teacher Education (PHZH), CH
13:00	Lunch break
14:00	Learning About AI by Creating an Own AI Project The Swiss AI Competition Fiona Könz ETH Zurich, CH
15:00	Ethel An AI-Based Virtual Teaching Assistant Gerd Kortemeyer ETH Zurich, CH
16:00	Coffee break
16:15	The Role of AI-Generated Explanations and Explainability in Education Panel Discussion Moderated by Francesco Sovrano Gerd Kortemeyer ETH Zurich, CH Rita Sevastjanova ETH Zurich, CH Luzia Sauer Zurich University of Teacher Education (PHZH), CH Sascha Schneider University of Zurich, CH
17:15	Closing remarks

Speakers' bios

Charlotte Axelsson

Head of Digital Learning
Zurich University of Teacher Education (PHZH), CH

With an interdisciplinary background in communication design and pedagogy, Charlotte Axelsson brings a visionary approach to digital learning, viewing it as a dynamic process that fosters creativity and curiosity. In her role, Axelsson leads initiatives that support students and educators in integrating digital media into teaching and learning. Her team provides resources and guidance to enhance digital competencies within the educational community. Axelsson is also involved in the “LeLa – Learning Lab Higher Education Didactics for Digital Skills” project, a collaboration among Zurich universities aimed at promoting digital skills among university teachers. This initiative focuses on innovation in higher education didactics and the development of digital competencies for educators in the digital age. Her work emphasizes the importance of viewing digital transformation in education not merely as the adoption of new tools but as a comprehensive, agile process that activates creativity and curiosity among learners and educators alike.

Sascha Schneider

Professor of Educational Technology
University of Zurich, CH

Sascha Schneider’s research focuses on the question of how learning media can be designed to improve learning. Methodologically, experiments are conducted and analyzed or data are meta-analytically evaluated. His research focuses on cognitive, emotional, motivational, social, and metacognitive processes in learning with digital media such as text- and image-based websites, animations, videos, but also interactive media such as learning video games and augmented and virtual reality environments. However, the focus is not only on teaching-learning processes, but also on retrieval processes that become necessary when working on learning tasks but also quizzes. In his dissertation he investigated the effects of decorative images on learning with media. In his postdoctoral thesis, he focused on the effect of choice as an autonomy enhancer and motivator in digital learning environments.

Rita Sevastjanova

Senior Researcher at the Interactive Visual Analytics Lab
ETH Zurich, CH

Rita Sevastjanova is a postdoctoral researcher in the Interactive Visualization and Intelligence Augmentation (IVIA) lab at ETH Zurich, with a focus on developing interactive visual methods for explaining large language models (LLMs). She is particularly interested in exploring the model representations of words (i.e., so-called contextual word embeddings) and their encoded properties, ranging from specific linguistic knowledge to harmful social biases.

Jaromir Šavelka

Research Associate in the Computer Science Department
Carnegie Mellon University, US

Jaromír Šavelka’s expertise lies at the intersection of artificial intelligence (AI) and law, focusing on the application of natural language processing (NLP) and machine learning to legal texts. Šavelka has developed information retrieval technologies to automate the interpretation of statutory and regulatory terms, enhancing legal workflows. In addition to his work in AI and law, Jaromir Šavelka contributes to computing education. He has co-authored studies evaluating the effectiveness of large language models, such as GPT-4, in answering multiple-choice questions about code, highlighting their potential and limitations in educational settings. His research also includes developing tools like CodeHelp, which leverages large language models to provide scalable support in programming classes. Šavelka’s interdisciplinary expertise in AI, law, and computing education positions him as a valuable contributor to discussions that bridge these fields.

Tanja Käser

Assistant Professor in the School of Computer and Communication Sciences
Swiss Federal Institute of Technology Lausanne (EPFL), CH

Tanja Käser leads the Machine Learning for Education Laboratory (ML4ED), where her research focuses on the intersection of machine learning, data mining, and education. She is particularly interested in creating accurate models of human behavior and learning. Prior to joining EPFL, Tanja Käser was a senior data scientist with the Swiss Data Science Center at ETH Zurich. Before that, she was a postdoctoral researcher with the AAALab at the Graduate School of Education of Stanford University (US). Tanja Käser received her doctoral degree from the Computer Science Department of ETH Zurich. In her dissertation, completed at the Computer Graphics Laboratory, she focused on user modeling and data mining in education, which was honored with the Fritz Kutter Award 2015.

Fiona Könz

Biochemist and Artist
ETH AI Center, CH

Fiona Könz holds a master’s degree from ETH Zurich in biochemistry and a BA in fine arts from the Zurich University of the Arts. She worked for five years as a freelance artist with her art partner Gregor Vogel, mainly doing participative performance projects in public space. From 2019 to 2023 she was part of the Strategic Foresight Team in the Office of the President at ETH Zurich, working on projects concerning the long-term future of universities and education. After one year at digitalswitzerland, she joined the ETH AI Center to lead the AI Challenge (formerly called KI Wettbewerb). The project offers teenagers a platform to actively learn about AI by creating their own AI projects. Her interests are at the intersection of public space, institutions, technology, and education.

Gerd Kortemeyer

Staff of Rector's Staff
ETH Zurich, CH

Gerd Kortemeyer is a member of the rectorate of ETH Zurich and an associate at the ETH AI Center. He is also an associate professor emeritus at Michigan State University (US). He holds a doctorate in physics from Michigan State University (US), where he taught for several decades. His research focusses on technology-enhanced learning of STEM disciplines; currently, he is advancing the research and development of AI-based tools and workflows for teaching, learning, and assessment.

Luzia Sauer

Lecturerer
Zurich University of Teacher Education (PHZH), CH

Luzia Sauer holds a PhD in applied linguistics from the University of Auckland (NZ), specializing in spoken second language acquisition (SLA), language learning motivation, and contexts for language learning. She is currently a lecturer at the Zurich University of Teacher Education (PHZH), where she teaches English methodology and language proficiency to future secondary school English teachers, including coaching sessions conducted in schools. Luzia Sauer’s research focuses on practical applications, such as developing EdTech solutions for English language learning in Swiss schools (e.g., a chatbot) and visualization tools for educational topics. Her extensive experience with English as a Foreign Language (EFL) practices stems from her work as a lecturer, researcher, and former vocational school teacher.

Margaritha Windisch

PhD Candidate
Center for Law & Economics, ETH Zurich, CH

Margaritha Windisch holds a Bachelor of Law in Business Law from the Vienna University of Economics and Business (AT) and a Master of Law in International and Comparative Law from the University of Lausanne (CH). In her research, she uses empirical methods to explore legally relevant questions of copyright authorship and infringement and examines underlying psychological aspects. Several projects focus on how technological developments, like generative AI, shape copyright law, particularly in visual artworks. Before joining ETH, Margaritha advanced digital education at Austrian schools and the University of Lausanne through various activities.

Abraham Bernstein

Professor in the Department of Informatics
Head of the Digital Society Initiative
University of Zurich, CH

Abraham Bernstein’s current research focuses on various aspects of the semantic web, data mining/machine learning, and collective intelligence. His work integrates social sciences, including organizational psychology, sociology, and economics, with technological and engineering fields, such as computer science and artificial intelligence. Before joining the University of Zurich, Abraham Bernstein was a faculty member at New York University (US) and also gained experience in the industry. He holds a doctoral degree from MIT (US) and a diploma in Computer Science from ETH Zurich.

Francesco Sovrano

Early-Career Fellow
Collegium Helveticum, CH

Francesco is a computer scientist and a data science researcher specializing in responsible AI. His work centres on explanations, examining their theoretical foundations and their role in enhancing transparency, learning, and information acquisition. Explanations are the core theme of his research. Francesco was awarded a PhD scholarship by the University of Bologna (IT) with an open research topic, which he dedicated to developing a computational theory of explanations. This work led to the creation of LLM-based intelligent user interfaces that improved textbook explanations and facilitated AI-driven software explanations for legal compliance. He also applied his theory to explain regulations to off-policy reinforcement learning agents, enhancing their learning efficiency. During his postdoctoral research at the University of Zurich, Francesco applied theories of explanation to software development and regulatory compliance within the EU, while also expanding into machine learning for code. His interest in understanding how LLMs generate outputs earned him an early-career fellowship at the Collegium Helveticum, where he is developing new explainable AI tools to uncover the rules and biases in LLM-generated explanations.