Abstract:
This study investigates the application of Artificial Intelligence (AI) in fraud
prevention and detection within public sector institutions in South-South Nigeria, grounded in
theories from computer science, information systems, and organizational management. Key
theories include information processing theory, bounded rationality, organizational learning,
and the fraud triangle theory. A mixed-methods approach, combining qualitative and
quantitative research methods, was employed to provide a comprehensive analysis. The study
focused on public sector employees, auditors, fraud investigators, and IT professionals across
six states: Rivers, Bayelsa, Delta, Akwa Ibom, Cross River, and Edo. A sample size of 300
respondents was determined using Krejcie and Morgan's table for sampling. Data collection
involved structured questionnaires to gather quantitative data on AI awareness and its
perceived effectiveness in fraud prevention and detection, and semi-structured interviews with
key informants to gain qualitative insights. Descriptive and inferential statistics were used to
analyze the data, including frequency distributions, mean, standard deviation, and regression
analysis using SPSS. The regression analysis revealed that Data Analytics, Machine Learning,
and Natural Language Processing significantly impact internal control and auditing, as well as
whistleblower programs. Data Analytics showed the strongest positive relationship with
internal control and auditing (B = 0.40, p < 0.001), followed by Machine Learning (B = 0.35, p
= 0.004), and Natural Language Processing (B = 0.25, p = 0.024). For whistleblower
programs, Machine Learning had the most substantial impact (B = 0.40, p = 0.008), followed
by Data Analytics (B = 0.30, p = 0.036). The findings underscore the critical role of AI
technologies in enhancing fraud prevention and detection mechanisms in the public sector.
Recommendations include prioritizing AI integration, investing in data analytics, developing
machine learning models, exploring NLP applications, and implementing comprehensive AI
training programs.