The ethics and 3R principle of experimental animals are basic tenets that should be strictly obeyed. However, because of the inner complexity of organisms, the data obtained in vivo tend to be gathered in voluminous amounts. There is a lack of appropriate method to thoroughly explore these datasets, which are too complicated and contain too much hidden information to be analyzed effectively. Therefore, the number of experimental animals performed is increasing. However, in the context of big data, and with the rapid development of computer technology, artificial intelligence (AI) has made splendid progress. Nowadays, AI is widely applied in the establishment of relevant databases and mathematical models and the identification and prediction of features of, e. g. , microscopic molecules, macroscale images, behavior, and basic physiological indicators, and it serves as a powerful and convenient auxiliary tool for scientific researchers. Compared with micro-analysis, the macroscopic application of character recognition seems more feasible and practicable. This paper summarizes the applications, challenges, and prospects of AI use in the recognition of phenomena hidden in experimental animal data.