Two student researchers with the U. S. National Science Foundation AI Institute for Adult Learning and Online Education (AI-ALOE) were selected for the 2025 AIVO AI4Ed Summer Research Fellowship, funded by the NSF and Google.org. The fellowship supports graduate researchers advancing artificial intelligence innovations in education.
Georgia Institute of Technology’s Joon Kum and Georgia State University’s Lia Haddadian represented AI-ALOE among a national cohort of fellows developing AI-driven learning tools. Their projects addressed key challenges in collaborative reasoning and professional learning for educators.
Advancing Collaborative Reasoning Through AI
Kum collaborated with a multi-university team on a project designed to make AI a more natural and effective collaborator in group learning environments. Their project, “Probing Minds: Modeling Implicit Disagreement and Designing Responsive AI to Support Collaborative Reasoning,” explored how AI can better interpret subtle cues, such as hedging or indirect language, that indicate disagreement during conversation.
“Collaborative learning thrives on moments of disagreement, but such moments are often conveyed subtly,” the team noted in their presentation abstract. By training and evaluating disagreement-detection models, including large language models, the researchers explored a socially aware AI agent capable of recognizing implicit signals and prompting deeper reasoning in small group discussions.
Personalized Learning for Speech-Language Pathologists
Haddadian’s team focused on professional learning through AI driven personalization. Their project, “AI-Augmented Personalized Podcast Delivery to Advance Language-Science Competence for Speech-Language Pathologists,” aimed to strengthen the language-science knowledge of speech-language pathologists (SLPs) by using AI to personalize professional learning through podcasts..
The team designed a demo platform that delivers podcast episodes tailored to the listener’s specific questions and learning needs. Using feedback from informal focus groups with SLP students, the researchers aligned the tool’s content and delivery with adult learning science theory to maximize engagement and relevance.
Showcasing Innovation at the AIVO AI4Ed Summit
Both projects were presented at the AIVO AI4Ed Summer Summit in Chicago, where fellows shared their outcomes and prototypes with peers, mentors, and representatives from NSF and Google.org.
Through their fellowship experiences, Kum and Haddadian demonstrated AI-ALOE’s commitment to advancing research at the intersection of AI and education—designing intelligent systems that enhance collaboration, personalization, and professional learning across diverse educational settings.
