AI Integration in Entrepreneurship Education: A Systematic Review of the Mediating Role of Self-Efficacy

The integration of artificial intelligence (AI) in entrepreneurship education has expanded rapidly following the launch of ChatGPT however, empirical evidence regarding its effectiveness and implementation challenges remains fragmented. This study synthesizes empirical evidence on AI integration in entrepreneurship education through a systematic literature review following PRISMA 2020 guidelines. A Scopus database search yielded 186 articles, with 25 empirical studies meeting inclusion criteria. Findings reveal neural networks and generative AI as dominant technologies, with neural networks applied in nearly half of the studies and generative AI in more than a quarter. Experiential learning emerged as the most effective pedagogical approach. Entrepreneurial self-efficacy was confirmed as a key mediator between AI and entrepreneurial intention, with consistently strong path coefficients and large effect sizes across studies. AI integration consistently enhanced entrepreneurial intention, self-efficacy, and knowledge acquisition. However, a notable decline in self-perceived creativity indicates potential deskilling risk. Primary challenges include limited technical infrastructure, faculty readiness, and ethical concerns. The concentration of studies in China constrains cross-cultural generalizability. This study proposes the Integrated AI-Entrepreneurship Education Model (IAEEM) as a conceptual framework and provides practical recommendations for higher education institutions to optimize AI integration while mitigating potential risks in developing students' entrepreneurial competencies effectively.