Assessing the Efficacy of AI-based Techniques in Anomaly Detection in Financial Institutions

As financial crimes grow in complexity, the adoption of Artificial Intelligence (AI) in financial institutions has proven to be a good instrument in safeguarding financial institutions against anomalies and fraudulent activities, though its adoption in developing countries is still questionable due to the high cyber theft rate. This paper therefore focused on ascertaining the Efficacy of AI-based techniques in Anomaly Detection in financial Institutions in Enugu State, Nigeria. The study utilized descriptive survey design, a quantitative base method focusing on banking sectors and universities in Enugu State. 108 professionals and stakeholders in financial, IT firm and lectures were the population used for the study. Total sampling techniques were adopted due to the manageable size of the population. A reliability index of 0.85 was established using Cronbach alpha to ascertain the internal consistency of the instrument. The research assistants involved in administering the instruments were briefed by the researcher. Data was collected through a structured questionnaire designed to capture quantitative responses from the respondent using both Google Forms and physical distribution. Data collected were analyzed using mean (x̄) and standard Deviation (σ) with the aid of Statistical Product and Service Solutions (SPSS) Version 26.The study found among others that Artificial Intelligent no doubt improves the accuracy of Know Your Customer (KYC) procedures, improves language processing for enhancing communication and transaction monitoring in financial institution. The study also found out that financial institution in Enugu State, Nigeria are faced with numerous challenges in effective adopting AI base techniques anomaly detections which are low technical knowhow; lack of collaborations among experts, low integration of Explainable AI (XAI) techniques, lack of consideration of regional and institutional differences in fraud behavior, high cost of procurement and maintaining the AI detection techniques among others. It was therefore recommended among others that increase in collaboration between data scientists and financial security experts, both local and international, will help in continually enhancing the effectiveness of fraud detection systems in financial institutions as well as attract supports.