Abstract: [Purpose/Significance] This study aims to investigate how to effectively utilize generative AI search engines combined with human-machine collaboration to expand scientific research topic selection, in order to address the challenges of information overload and difficulties in selecting research topics. [Design/Methodology] This research proposes a topic selection workflow based on generative AI search engines in response to the current research landscape, and verifies the topic selection process through experimental and interview methods. The study discusses the advantages of the "human-in-the-loop" in the topic selection process based on the results. [Findings/Conclusion] In the experiment, all participants give a satisfaction score of 90 or above（out of 100）for the effect of topic expansion. Generative AI search engines can effectively assist in scientific research topic selection, but their maximum utility can only be achieved by adopting a human-machine collaborative approach. The "human-in-the-loop" paradigm can fully leverage the strengths of AI, enhancing the quality and efficiency of topic selection and propelling scientific research to a higher level of development. [Originality/Value] The method of expanding topic selection based on generative AI search engines and human-machine integration offers researchers a novel way for topic selection. This method aids in alleviating the issues of information overload and the difficulties associated with selecting research topics, providing theoretical guidance for researchers.

Keywords: Generative AI, Search engine, Human-computer collaboration, Research topic selection