This study examines the interaction of community college students with ChatGPT during computational problem-solving tasks in Excel, focusing on usage patterns, perceptions, and the influence of instructor guidance. As ChatGPT becomes increasingly integrated into both educational and professional settings, it is critical to understand its impact on students’ task value and motivation. Grounded by Expectancy-Value Theory, this study employs a mixed-methods approach, combining quantitative data (perceived task value and task motivation) with qualitative data (reflective reports and screen recordings) analyzed through thematic coding and Epistemic Network Analysis (ENA). The research explores changes in students’ perceptions and usage patterns of ChatGPT across two sessions that differed in the presence of instructor guidance. Results reveal that instructor guidance enhanced students’ task motivation and perceived task value by helping them recognize ChatGPT's utility, thereby fostering its use as a collaborative learning partner. The study underscores the importance of designing instructional environments that promote ethical and effective use of generative AI, supporting student agency and deeper integration into educational practices. These insights offer valuable guidance for educators seeking to enhance student engagement through effective AI-supported learning experiences. 

Keywords : ChatGPT, Community college, Expectancy-value theory, Computational problem-solving, Epistemic Network Analysis