Applications of Multimodal Large Language Models in the Accounting Industry

This paper explores the transformative potential of multimodal large language models (MLLMs) in the accounting industry, examining how the integration of text, images, and numerical data can enhance financial analysis, reporting, and decision-making processes. Through a comprehensive literature review and theoretical analysis, the study establishes a foundation for understanding the technical and operational capabilities of MLLMs in accounting contexts. An empirical research design is employed to evaluate real-world applications, including automated financial reporting, anomaly detection, and document classification, demonstrating significant improvements in accuracy, efficiency, and scalability compared to traditional methods. The findings reveal both opportunities and challenges, such as data privacy concerns, model interpretability, and user adoption barriers. Accounting professionals' feedback underscores the technology's promise while highlighting the need for improved integration with existing systems and ethical frameworks. The study concludes by emphasizing the importance of interdisciplinary collaboration and ongoing innovation to fully leverage MLLMs in finance. Future research directions include refining model robustness, developing AI governance standards, and expanding applications across broader financial domains.