Comparing Student Acceptance of ChatGPT and Claude AI Through an Extended Technology Acceptance Model

Abstract

The classical Technology Acceptance Model falls short in fully explaining how students embrace generative artificial intelligence, as the impact of trust on their intention to use varies by platform. This study explores the acceptance mechanisms of ChatGPT and Claude AI by extending the Technology Acceptance Model to include Trust and Prior Experience. This study involved 312 students from the Faculty of Science and Technology at Universitas Muhammadiyah Tapanuli Selatan, selected through purposive sampling. Data were gathered using a validated Likert-scale questionnaire and analyzed with Partial Least Squares Structural Equation Modeling, incorporating multi-group analysis. Both measurement models demonstrated satisfactory validity and reliability. For ChatGPT, perceived usefulness was the primary factor influencing attitude, but this did not lead to a corresponding behavioral intention. In contrast, for Claude AI, both trust and attitude significantly influenced behavioral intention. Among the eight structural paths, four showed significant differences between platforms, with the extended model providing a better explanation for Claude AI. These results suggest that the acceptance of generative AI is dependent on the platform, offering conceptual insights into the structural assumptions of the TAM and practical guidance for platform-specific AI integration policies in higher education.