TY - CHAP

T1 - From Feedback to Formative Guidance

T2 - 33rd Conference on User Modeling, Adaptation and Personalization - UMAP 2025

AU - Ghoochani, Fatemeh

AU - Scharfenberger, Jonas

AU - Funk, Burkhardt

AU - Doublan, Raoul

AU - Jakharabhai Odedra, Mayur

AU - Etsiwah, Bennet

N1 - Conference code: 33

PY - 2025/6/12

Y1 - 2025/6/12

N2 - Large Language Models (LLMs) offer scalable opportunities to personalize feedback in education, yet their trustworthiness and effectiveness remain underexplored. We present a study conducted in an introductory programming and data science course with approximately 1,400 first-year university students. A subset of these students received both peer and LLM-generated feedback on their individual programming projects. Our results show that 56% of students preferred the LLM feedback, and 52% could not reliably distinguish it from human-written feedback. Student ratings suggest that LLM feedback is perceived as helpful, constructive, and relevant, though it often lacks personalized depth and motivational nuance. These findings underline the potential of LLMs to support scalable, personalized education, while pointing to key areas for responsible improvement. Based on these insights, we outline the future roadmap for the course in which LLM-generated feedback supports students in their learning journey but also instructors through monitoring student performance and helping to allocate instructional resources more effectively. Given limited human resources this approach enables personalized instructor feedback to be scaled to a large group of students.

AB - Large Language Models (LLMs) offer scalable opportunities to personalize feedback in education, yet their trustworthiness and effectiveness remain underexplored. We present a study conducted in an introductory programming and data science course with approximately 1,400 first-year university students. A subset of these students received both peer and LLM-generated feedback on their individual programming projects. Our results show that 56% of students preferred the LLM feedback, and 52% could not reliably distinguish it from human-written feedback. Student ratings suggest that LLM feedback is perceived as helpful, constructive, and relevant, though it often lacks personalized depth and motivational nuance. These findings underline the potential of LLMs to support scalable, personalized education, while pointing to key areas for responsible improvement. Based on these insights, we outline the future roadmap for the course in which LLM-generated feedback supports students in their learning journey but also instructors through monitoring student performance and helping to allocate instructional resources more effectively. Given limited human resources this approach enables personalized instructor feedback to be scaled to a large group of students.

KW - artificial intelligence

KW - education

KW - large language models

KW - personalized feedback

KW - programming

KW - Informatics

UR - http://www.scopus.com/inward/record.url?scp=105011049357&partnerID=8YFLogxK

U2 - 10.1145/3708319.3733808

DO - 10.1145/3708319.3733808

M3 - Article in conference proceedings

AN - SCOPUS:105011049357

T3 - UMAP 2025 - Adjunct Proceedings of the 33rd ACM Conference on User Modeling, Adaptation and Personalization

SP - 398

EP - 403

BT - UMAP 2025 - Adjunct Proceedings of the 33rd ACM Conference on User Modeling, Adaptation and Personalization

A2 - Conati, Cristina

A2 - Narducci, Fedelucio

A2 - Rossiello, Gaetano

A2 - Musto, Cataldo

A2 - Vassileva, Julita

PB - Association for Computing Machinery, Inc

Y2 - 16 June 2025 through 19 June 2025

ER -