Cultural-Historical Activity Theory and AI: Innovating and Optimizing Financial Data Retrieval

Abstract

Machine learning is a transformative technology with profound applications across various domains, including finance. This study introduces the Artificial Intelligent Financial Model (AIFM), developed to enhance the accuracy and efficiency of retrieving CEO pay and median employee pay. Despite significant advances in chatbot technologies within financial services, accurate evaluation of these systems, particularly in terms of handling complex financial documents, remains challenging. Traditional models often fail to capture the complexities of financial data analysis through effective, prompt engineering. To address these limitations, the AIFM uses advanced machine learning and NLP to evaluate the system’s response to simulated queries. This involves sophisticated prompt engineering strategies that ensure precise and reliable evaluation of the chatbot’s performance, enabling the system to process inputs accurately without the need for real-time human feedback. This approach allows for controlled testing and refinement of the chatbot’s capabilities in a consistent and repeatable environment. Moreover, this research incorporates Cultural-Historical Activity Theory (CHAT) from a technical perspective, deviating from the typical qualitative approaches often seen in other papers. We aim to provide a structured and measurable analysis of the AI’s interaction logic and its ability to process and present financial data accurately. Our findings demonstrate the effectiveness of the AIFM in providing detailed and accessible financial insights, which could significantly impact on the broader field of finance by introducing new levels of precision and analysis-focused interaction. This study proposes a novel approach that could expand the current understanding of AI’s capability in high-stakes environments.