Project-Based Learning (PBL) is an educational approach rooted in constructivist principles. It focuses on addressing real-world issues and promotes learning through collaborative exploration and iterative thought among learners. This method shifts the educational focus from simply memorizing facts to developing and enhancing skills. Nevertheless, PBL encounters several obstacles in actual teaching practices, particularly in three areas: tailored project design, ongoing process guidance, and sophisticated evaluation feedback. As digital transformation in education accelerates and large-scale model technology advances, the incorporation of Generative AI into PBL offers support for its effective execution. Generative AI can aid in refining project design for teaching, facilitate personalized exploration processes for high school students, and enable comprehensive skill assessments for secondary school learners. Currently, the use of Generative AI in PBL is still in a research phase. This research focuses on creating a design that emphasizes the value of artificial intelligence in enhancing the quality of learning, particularly when utilizing active methodologies such as project-based learning (PBL). The study examined teachers' perceptions of PBL that integrates AI compared to those that do not in the context of secondary higher education. A sample of 385 secondary school teachers participated in this research, which explored perceived efficacy, the personalization of learning through AI, and levels of motivation. This article breaks down the use of Generative AI in the phases of project design, execution, and assessment. Using the teaching project of constructing an exchange network as a case study, this framework suggests a method for prompt language design of Generative AI in PBL project development. Lastly, it assesses the potential risks associated with Generative AI. This research offers a framework and example for intelligent innovation within PBL.