Abstract: Since the 21st century, artificial intelligence (AI) technologies, particularly those exemplified by deep learning, have significantly accelerated the interdisciplinary integration of various fields. This integration has led to a surge of AI-driven scientific research, especially in the application of AI within the realm of biology (AI4B). Over the past seven decades, AI and biology have mutually influenced one another; while AI has enabled the precise resolution of complex biological problems, it has simultaneously fostered its own continuous iteration and advancement. AI4B has established a transformative model for the production of scientific knowledge through human-machine collaboration, which can be delineated across three dimensions: in the tool empowerment dimension, AI enhances task efficiency; in the deep embedding dimension, AI reshapes the scientific innovation chain; and in the general platform dimension, AI addresses multimodal tasks. The development of a human-machine symbiotic AI4B knowledge ecosystem is of paramount importance for effectively mitigating AI-related risks, formulating pertinent policies, and shaping the "emerging science and technology" that AI represents.