This paper summarizes the common practice patterns of artificial intelligence application in criminal investigation, mainly including video investigation, individual identification, data retrieval, information mining, auxiliary interrogation, evidence evaluation, etc., which brings multi-dimensional lock, situation and finger integration, and full chain support to the investigation department. Then this paper analyzes a series of risks and challenges accompanied by the application of artificial intelligence in criminal investigation, such as black box effect, precision deviation, algorithm overstep, open panorama, investigation secret room and so on. Some suggestions are put forward, such as anchoring the basic stance of human-oriented and human-computer integration, promoting the integration of data-driven and knowledge-guided methods, enhancing the ethical adaptation of artificial intelligence application in criminal investigation, and optimizing the legal system of artificial intelligence application in criminal investigation.

artificial intelligence  /  investigation  /  technical effect  /  human-machine integration  /  knowledge to guide