Vibe Coding and AI-assisted Programming in Undergraduate Computer Education: A Systematic Review of Pedagogical Benefits and Foundational Skill Risks

Abstract

This study aimed to identify and synthesize how vibe coding and AI-assisted programming are defined, conceptualized, and operationalized in undergraduate computer education programs; examine and categorize the reported pedagogical benefits of AI-assisted coding approaches; and identify the risks and challenges associated with them. A systematic review was conducted across major academic databases, including Scopus, IEEE Xplore, ACM Digital Library, and ERIC, covering the period from 2023 to 2025. A total of 19 studies met the predefined inclusion criteria for analysis. The synthesized findings reveal that GitHub Copilot and ChatGPT were often used as an AI pair programmer and an AI conversational assistant, respectively. The primary pedagogical benefits are the acceleration of rapid prototyping, the reduction of cognitive load associated with syntactic errors, and increased student motivation. Students often transition from low-level to higher-level coding, focusing on problem definition and critical code review. Conversely, significant foundational skill risks were identified, including a potential decline in core debugging skills, reduced deep understanding of algorithms and data structures, and an increased reliance on generated code, which may hinder mastery of fundamental programming constructs. The review also highlighted the emerging challenge of teaching proper prompt engineering and adequate code verification as essential new skills.