Development and Implementation of an AI Tutor To Scaffold Problem-Solving in Introductory Chemistry

Abstract

Developing problem-solving competency is a central goal of undergraduate chemistry education, yet many students struggle to engage productively with complex, multistep problems. At the same time, the rapid adoption of large language model (LLM)–based tools by students in STEM courses raises concerns about over-reliance and uncritical use, potentially undermining learning. In this work, we present the design, implementation, and evaluation of STEPS (STEM Tutor for Effective Problem Solving), an AI tutor particularly designed to guide students in creating a problem-solving plan without divulging solutions. The tutor was deployed in an introductory chemistry course at an R1 university, and its impact was examined using performance data, Likert-scale survey responses, and qualitative analysis of student reflections. Students demonstrated high performance on STEPS-assisted problems and reported strong confidence in their solutions. Survey results indicated consistently high perceptions of helpfulness. Qualitative analysis revealed that students valued the tutor’s guidance in organizing their thinking and checking reasoning without giving away solutions. These findings suggest that carefully designed, task-specific AI tutors can promote productive engagement with problem solving while mitigating risks associated with uncritical AI use in chemistry education.