Large language models in the banking sector: Areas, opportunities, and threats

Abstract

Large language models (LLMs), a rapidly growing area of generative AI (GenAI), are being increasingly applied in banks. The article aims to identify the areas in which banks are using LLMs and to analyse the opportunities and threats associated with their implementation from the perspective of both banking institutions and their customers. The article employs a critical review of the relevant literature and a comparative method. Statistical secondary data from sources such as the Statista database and reports from Accenture, KPMG, PwC, and the European Banking Authority (EBA) support the research. The findings indicate that banks are primarily utilizing LLMs in customer service and internal operations, such as risk management, thereby contributing to their digital transformation. Banks seem to experience more benefits from the implementation of LLMs than their customers. The main challenges identified include legal, ethical, and customer acceptance issues.