Abstract This study examined the relationship between artificial intelligence and tax administration in Nigeria. A quantitative research design was employed, using a structured questionnaire administered to a purposively selected sample of 552 tax administrators across three states in Nigeria. Of the 232 questionnaires distributed to Delta, Bayelsa, and Rivers States, 202 were returned, reflecting a high response rate of 87.07 percent. Hypotheses were tested using the Pearson Product Moment Correlation Coefficient and the multiple regression analysis technique. Results revealed that Machine Learning has a significant, moderate positive relationship with data collection (r = .521, p < .001), leading to the rejection of the first null hypothesis. However, its relationship with data processing was not statistically significant (r = .092, p = .184), resulting in acceptance of the second null hypothesis. Furthermore, Natural Language Processing showed a significant, moderate positive relationship with both data collection (r = .547, p < .001) and data processing (r = .532, p < .001), leading to the rejection of the third and fourth null hypotheses. Based on these findings, the study recommends that tax authorities strategically adopt Machine Learning to automate and enhance the accuracy of data collection processes, while integrating Natural Language Processing to improve the analysis of unstructured taxpayer communications and documents. The synergistic implementation of these technologies, supported by continuous staff training, is essential for optimizing administrative efficiency and revenue mobilization.
Titcombe et al. (2026) studied this question.