AI-based learning assistants are gaining popularity at a rapid pace. This research is dedicated to establishing empirically validated design principles for the effective development and implementation of AI assistants in educational settings. Our primary focus is on ensuring learner acceptance and accommodating their preferences. How can AI assistants be designed to enhance learner acceptance and cater to their preferences in educational environments? Rooted in the KIAM model, this research examines AI acceptance, emphasizing the alignment between technology and learner needs. The methodology includes qualitative and quantitative approaches – surveys and focus groups with learners to assess attitudes, behaviors, and preferences. Experimental studies observe interactions between learners and prototype AI assistants, providing data on usability and effectiveness. Key findings highlight transparency, trust, and fairness as crucial components driving technology acceptance. Based on focus groups (n = 24), the study proposes AI assistant design principles such as user-friendly interfaces, consistent design, adaptability to individual learner needs, customization options, and emotional intelligence of the learning companion. In this context, emotional intelligence entails the AI assistant responding empathetically to students' emotional expressions and offering constructive guidance. Additionally, the importance of integrating gamification for motivation is emphasized. These principles aim to enhance aesthetic appeal, support functionality, and ensure a positive learning experience. A personalized AI assistant considers learning pace, rhythm, and progress tracking for targeted interventions. These empirically grounded design principles offer a comprehensive framework for developers and educators to create AI assistants that align with user needs, promoting acceptance, and fostering positive learning outcomes.

Authors:
Sandra Hummel, ScaDS.AI; Dresden University of Technology, Germany
Mana-Teresa Donner, ScaDS.AI; Dresden University of Technology, Germany

About the Presenter(s)
Mana-Teresa Donner is a psychologist and educational scientist specializing in learning analytics, higher education pedagogy, and self-regulated learning. As a PhD candidate, Mana-Teresa is exploring the learning psychology aspects of AI-based mentoring.
https://www.researchgate.net/profile/Mana_Teresa_Donner2

Sandra Hummel, an educational scientist and research group leader at ScaDS.AI at TUD, focuses on human-centered AI-based teaching and learning. She coordinates several EU projects developing AI tools tailored to the needs of diverse learners.
https://www.researchgate.net/profile/Sandra-Hummel-4

Connect on Linkedin
http://www.linkedin.com/in/mana-teresa-donner-9ba1aa217

Connect on ResearchGate
https://www.researchgate.net/profile/Mana_Teresa_Donner2

See this presentation on the full schedule – Saturday Schedule

Bookmark