Schlagwort: AI

Veröffentlicht am

[podcast] KI-gestützte Lehre #tugraz

Quelle

Es freut uns, dass wir nun den 25. Teil unserer Podcast-Serie „Lehren – Lernen – Lauschen“ auf der TELucation-Webseite zur Verfügung stellen können. Diesesmal spricht Benedikt Brünner zu „KI-gestützte Lehre„:

In dieser Folge spricht Benedikt Brünner vom Institute of Human-Centred Computing über seine Forschung zu digitaler Bildung und den verantwortungsvollen Einsatz von KI im Lehr- und Lernprozess. Er gibt Einblicke in seinen Weg vom Lehramt zurück an die TU Graz, erläutert Chancen und Grenzen von KI in der Hochschullehre und zeigt, wie KI-Tutor:innen, gezieltes Prompting und KI-gestützte Lernmaterialien Lehrende und Studierende unterstützen können.

[#27 – Benedikt Brünner zu KI-gestützte Lehre]

Und nicht übersehen – der Podcast ist auch in allen gängigen Portalen verfügbar:

Damit möchte ich mich auch in die Weihnachtspause 2025 verabschieden und mich gleichzeitig bedanken, dass so viele Personen hier mitlesen. Ich darf allen ein schönes Weihnachtsfest 2025 wünschen, sowie ein gutes und gesundes Jahr 2026.
Im Jänner 2026 beginn ich gerne wieder mit alter Frische.

Veröffentlicht am

[publication] Exploring genAI Chatbots in MOOCs: Analyzing Student Interactions and Self-Regulated Learning Behaviors #imoox #tugraz #research

Our publication about „Exploring genAI Chatbots in MOOCs: Analyzing Student Interactions and Self-Regulated Learning Behaviors“ got published.

Abstract:
The integration of generative AI (genAI) chatbots into Massive Open Online Courses (MOOCs) presents new opportunities for supporting self-regulated learning (SRL). This study examines 1,302 chat-bot interactions from two Austrian blended MOOCs, where a retrieval-augmented generation (RAG) chatbot based on GPT 4o-mini was deployed to assist students. Using the process-action framework by Lai (2024), we categorize chatbot interactions into key SRL processes: defining , seeking, engaging, and reflecting. Results show that students predominantly use the chatbot for information retrieval, content summarization, and quiz-based reinforcement, with 41% of interactions classified as information search queries and 17% as rehearsal. However, engagement with metacognitive SRL strategies, such as goal setting and self-evaluation, remains low. Additionally, non-learning interactions, including humor-driven conversations, functional queries, and prompt injection attempts, showcase ways students interact with AI in educational settings. Based on our findings, we propose refinements to the existing SRL process-action framework, incorporating new categories to better account for genAI chatbot-specific interactions, such as Evaluation of Information Quality and Reformatting. We discuss implications for chat-bot integration in MOOCs, emphasizing AI-generated quizzes, structured feedback, and safeguards against misuse.

[article @ book’s homepage]
[draft @ ResearchGate]

Rerence: Brünner, B., Ebner, M., Schön, S. (2025). Exploring GenAI Chatbots in MOOCs: Analyzing Student Interactions and Self-regulated Learning Behaviors. In: Hamonic, E., Sharrock, R. (eds) Digital Education: Shaping Sustainable Lifelong Learning for All in the Era of AI. EMOOCS 2025. Lecture Notes in Computer Science, vol 15733. Springer, Cham. https://doi.org/10.1007/978-3-032-00056-9_2

This is an impactful contributions, methodological rigor, and exceptional novelty in the research field of AI in education using a Chatbot within a MOOC-platform (iMooX.at)

Veröffentlicht am

[bericht] Ergebnisse aus der Fallberatung im KI-Lab Münster: Acht potentielle Booster-Anwendungen für die Hochschulbildung #KI

In Rahmen des KI-Labs im Sommer 2025 in Münster haben wir rund um mögliche sinnvolle KI-Anwendungen in der Hochschullehre diskutiert und diese dann gesammelt:

Wie lassen sich KI-Technologien künftig sinnvoll in Lehr- und Prüfungsprozesse einbinden?
Diese Frage stand im Mittelpunkt des ersten KI-Labs des HFD, das im Juli 2025 in Münster stattfand. Über 40 Expert:innen und Entscheider:innen aus Hochschulen, die sich strategisch mit dem Einsatz von Künstlicher Intelligenz befassen, arbeiteten dort gemeinsam an praxisnahen Lösungen. Martin Ebner und Sandra Schön von der TU Graz berichten, welche Ideen für KI-gestützte Lehr- und Lernanwendungen sie mit ihren Kolleginnen und Kollegen von anderen Hochschulen entwickelt haben. 

Zusammenfassend ist daraus dieser Bericht entstanden der jetzt vom Hochschulforum Digitalisierung veröffentlicht wurde: [Link zum Bericht]

Veröffentlicht am

[publication] Ensuring Quality in AI-Generated Multiple-Choice Questions for Higher Education with the QUEST Framework #AI #tugraz #research

This is an impactful contributions, methodological rigor, and exceptional novelty in the research field of AI in education.

Another research paper about the use of AI for education – titled „Ensuring Quality in AI-Generated Multiple-Choice Questions for Higher Education with the QUEST Framework“ – was published.

Abstract:
With the rise of generative AI models, such as large language models (LLMs), in educational settings, there is a growing demand to ensure the quality of AI-generated multiple-choice questions (MCQs) used in higher education. Traditional quiz development methods fall short in addressing the unique challenges posed by AI-generated content, such as consistency, cognitive demand, and question uniqueness. This paper presents the QUEST framework, a structured approach designed specifically to evaluate the quality of LLM-generated MCQs across five dimensions: Quality, Uniqueness, Effort, Structure, and Transparency. Following an iterative research process, AI-generated questions were assessed and refined using QUEST, revealing that the framework effectively improves question clarity, relevance, and educational value. The findings suggest that QUEST is a viable tool for educators to maintain high-quality standards in AI-generated assessments, ensuring these resources meet the pedagogical needs of diverse learners in higher education.

[draft @ ResearchGate]
[full article @ publisher’s website]

Reference: Ebner, M., Brünner, B., Forjan, N., Schön, S. (2025). Ensuring Quality in AI-Generated Multiple-Choice Questions for Higher Education with the QUEST Framework. In: Tomczyk, Ł. (eds) New Media Pedagogy: Research Trends, Methodological Challenges, and Successful Implementations. NMP 2024. Communications in Computer and Information Science, vol 2537. Springer, Cham. https://doi.org/10.1007/978-3-031-95627-0_20

Veröffentlicht am

[publication] aicast: Combining AI-Generated and Instructor-Defined Content in Educational Podcasts #research #tugraz

At this year’s ED-Media conference, we introduced a new AI-based concept for educational podcasts: „aicast: Combining AI-Generated and Instructor-Defined Content in Educational Podcasts„.

Abstract:
Generative AI is transforming educational content. In this poster, we present aicast.fyi, an open-source platform for producing personalized educational podcasts that combine AI-generated and instructor-defined content. Each track contains dynamic and fixed segments. Instructors define guiding questions during track design, ensuring alignment with learning objectives. This hybrid approach addresses concerns about maintaining instructional integrity with automation and personalization. The poster provides an architectural overview and demonstrates how the system integrates text generation and voice synthesis to produce audio-based learning experiences. Instructional designers must play an active role in shaping AI-assisted educational tools. This poster invites educators, designers, and researchers to engage with aicast and collaborate on developing high-quality, responsible, and learner-centered audio content. By allowing learners to control content sequences and personalize segments based on their own inputs, aicast.fyi also supports self-regulated learning (SRL), empowering users to take ownership of their educational journey.

[draft @ ResearchGate]
[full article @ conference website]

Reference: Brünner, B., Leitner, P., Geier, G. & Ebner, M. (2025). aicast: Combining AI-Generated and Instructor-Defined Content in Educational Podcasts. In T. Bastiaens (Ed.), Proceedings of EdMedia + Innovate Learning (pp. 148-152). Barcelona, Spain: Association for the Advancement of Computing in Education (AACE). Retrieved June 4, 2025 from https://www.learntechlib.org/primary/p/226139/

Veröffentlicht am

[presentation] Synthetic Educators: Analyzing AI-Driven Avatars in Digital Learning Environments #HCII25

Our research about AI-Driven Avatars was presented at 27th International Conference on Human-Computer Interaction, Gothenburg, Sweden.

Struger, P., Brünner, B., & Ebner, M. (2025, Juni 23). Presentation: Synthetic Educators: Analyzing AI-Driven Avatars in Digital Learning Environments. Graz University of Technology. https://doi.org/10.3217/bngt5-p2053

Veröffentlicht am

[publication] Workshop for Special Interest Group Emerging Technologies for Learning and Teaching: Exploring Educational Podcasts with AI #tugraz #research

We also did workshops at this year’s ED-Media conference. One of them was titled „Workshop for Special Interest Group Emerging Technologies for Learning and Teaching: Exploring Educational Podcasts with AI“ and followed the idea of open discussions.

Abstract:
How can educators unleash the power of AI without losing control over core educational strategies? In this workshop, we introduced aicast, an open-source platform for educational podcasts that combines the best of both worlds: AI-generated elements for personalization and flexibility, and fixed elements, with instructor-defined content to ensure pedagogical accuracy. This hybrid approach reduces the risks associated with AI-generated materials like hallucinations. Attendees experienced how the platform utilizes LLMs like ChatGPT for personalized content authoring and ElevenLabs for high-quality voice synthesis, enabling real-time creation of educational audio content. After a short demo and hands-on session, participants engaged in a guided discussion: What is the core of an educational podcast? Most importantly from the perspective of an instructional designer, how must an educational podcast be? This session was part of the Special Interest Group on Emerging Technologies for Learning and Teaching at the ED-Media 2025 conference.

[draft @ ResearchGate]
[full article @ conference website]

Reference: Ebner, M. & Brünner, B. (2025). Workshop for Special Interest Group Emerging Technologies for Learning and Teaching: Exploring Educational Podcasts with AI. In T. Bastiaens (Ed.), Proceedings of EdMedia + Innovate Learning (pp. 1375-1376). Barcelona, Spain: Association for the Advancement of Computing in Education (AACE). Retrieved June 4, 2025 from https://www.learntechlib.org/primary/p/226347/.