Generative AI in Accounting Education: Student Perceptions and Effectiveness in an Advanced Accounting Course

This study develops and applies an integrated framework combining the Technology Acceptance Model (TAM) and Self-Determination Theory (SDT) to examine generative artificial intelligence’s (AI’s) role in enabling personalized learning within accounting pedagogy. It investigates student perceptions of AI’s perceived usefulness and ease of use alongside its capacity to satisfy psychological needs for autonomy, competence, and relatedness, while assessing effectiveness in enhancing learning efficiency, engagement, knowledge retention, and professional skill development in an Advanced Accounting course at a Sino-foreign joint university. A mixed-methods design was employed, utilizing an online survey with closed- and open-ended questions distributed to 100 students enrolled in an Advanced Accounting course. The instrument captured AI familiarity, usage frequency, preferred assistance formats, willingness to adopt personalized exercises, comparative feedback quality, and ethical concerns. Quantitative data were summarized descriptively (frequencies, percentages) and analyzed via cross-tabulations to identify patterns by academic year and usage level. Qualitative responses underwent thematic analysis to explore personalization mechanisms, benefits, limitations, and future expectations, interpreted through TAM (usefulness/ease of use) and SDT (autonomy/competence/relatedness) lenses. Generative AI exhibits strong adaptability for foundational accounting tasks, with 55% of students willing to use personalized exercises and 45% preferring step-by-step guidance (particularly freshmen). Daily users rate AI feedback comparably or superior to instructor input (60%), reporting up to 29% theoretical improvement and 34% faster skill mastery. However, effectiveness declines for complex judgment-based tasks (35% view AI as less effective). Significant concerns include data privacy (70%), overreliance reducing critical thinking (28%), accuracy demands (67% upper-year students), and equity issues from unequal access. Preferences vary by academic stage: lower-year students favor structured support, while graduates prioritize real-time correction. This research advances accounting pedagogy by integrating TAM and SDT within a Sino-foreign university context, demonstrating AI’s potential to operationalize the Zone of Proximal Development for technical mastery while revealing relational and ethical limitations. It introduces the AI Adaptation Layer Theory to categorize accounting knowledge for AI processing versus human judgment, offering evidence-based guidance for hybrid models that balance personalization with critical thinking preservation. The findings inform phased implementation, ethical safeguards, and stage-specific design, contributing to equitable, effective AI integration that prepares accounting students for professional demands amid technological transformation. © 2027 selection and editorial matter, Md Jahidur Rahman, Tarek Rana and Hongtao Zhu; individual chapters, the contributors.