Unveiling sentiments of the cullen commission: Exploring AML compliance and regulation through deep learning techniques

Abstract

This paper examines the role of anti-money laundering (AML) regulations and compliance in combating money laundering and terrorist financing (ML/TF) in Canada. AML regulations establish guidelines for financial institutions to identify, prevent, and report suspicious activities. However, tensions often arise between AML compliance and regulation due to the challenges associated with implementing and adhering to AML regulations. This paper aims to investigate the sentiments expressed by individuals involved in AML compliance and regulation regarding the effectiveness of AML measures in the Canadian financial sector. The paper uses advanced deep learning (DL) methods like convolutional neural networks (CNN), recurrent neural networks with long short-term memory (RNN+LSTM), and pre-trained models like GloVe and BERT to explore emotions in the context of AML compliance and regulation. The findings indicate that DL models excel at accurately classifying sentiments from testimonies related to AML compliance and regulation. However, there are challenges in accurately capturing negative sentiments, which reflect the complexities and nuances associated with expressing criticisms about regulatory standards. The study emphasizes the importance of understanding the interplay between rational decision-making in compliance and the inherent conflicts with regulation. This article also highlights how DL models can potentially enhance sentiment analysis in the AML enterprise, enabling analysts to make decisions and policies based on financial intelligence. Nevertheless, DL models are not always easy to comprehend. Further research is needed to enhance the understandability and scalability of DL models when analyzing different AML datasets.