Work in Progress: Investigating ChatGPT for Grading Algebra Problems

This work-in-progress paper investigates the integration of ChatGPT, an advanced AI language model, as a grading tool for algebra problems in educational settings. From an academic perspective, the study examines the model’s ability to assess a range of algebraic tasks, including solving equations, simplifying expressions, and evaluating word problems. To ensure applicability within a structured learning environment, a pilot implementation is proposed within an existing course, where students are divided into two groups: an Experimental Group, graded using ChatGPT-enhanced feedback, and a Control Group, assessed through traditional human grading. The study evaluates the accuracy, consistency, and feedback quality of AI-driven grading compared to conventional assessment methods. Key areas of analysis include the AI’s ability to provide partial credit for alternative solutions, interpret symbolic algebraic notation, and address common student errors. By applying structured evaluation instruments to both groups, the research assesses the feasibility of AI-supported grading in improving grading efficiency while maintaining assessment fairness. Findings from the pilot study will inform the potential for broader adoption of AI-driven assessment tools in mathematics education, ensuring alignment with pedagogical best practices and student learning outcomes.