English for General Academic Purpose (EGAP) is crucial for cultivating high-quality innovative talents, focusing on developing their practical competence in academic English for communication, research and professional practice. With the in-depth integration of artificial intelligence (AI) into education, traditional EGAP teaching faces challenges like insufficient interaction and inefficient personalized guidance. Thus, human-computer collaborative teaching becomes an inevitable direction for EGAP reform. Based on previous research deficiencies and the latest AIGC development, this study optimizes the “one-core, three-orders and seven-synergies” model (grounded in PBL and supported by AIGC), integrating strengths of teachers, AI systems and students to expand EGAP teaching boundaries. Supplemented with new empirical data and cases, it verifies the upgraded model’s effectiveness, providing targeted theoretical and practical support for AI-EGAP integration and promoting EGAP teaching quality.