<p>This paper aimed to analyze how the initial patent citation influenced the citation metrics of academic papers, focusing on a select group of 4,919 highly cited papers in the field of artificial intelligence, spanning from 1990 to 2012. These papers were the top 1% most cited papers in their field and were simultaneously indexed in both the Web of Science (WoS) and the Lens databases. Our dataset also included 1,480,971 citing papers sourced from WoS and 42,887 citing patents derived from the Lens databases. Initially, we established several indicators to characterize the patterns of patent and paper citations for these highly cited papers. Then we employed Savitzky-Golay filter to depict the monthly trend of citation changes and BERTopic to cluster the paper topics. The Two-Way Fixed Effects Model was performed to infer the causal impact of patent citation on the paper citation. Statistical results indicated that the citation difference between papers with patent citation and papers without patent citation did not significantly change over time prior to the patent citation. Furthermore, patent citation contributed to the increased academic impact and longer citation rising phase of highly cited papers in the field of artificial intelligence. After being cited by patents, highly cited papers were cited approximately 23 times more per year compared to those that were not cited by patents. Besides, researchers in the industrial community paid more attention to the academic papers after patent citation.</p>