The financial domain is losing ground to rapid-developing fraud schemes. It puts intense pressure on organizations such as banks to find new approaches to tackle financial misconduct. This financial crime has always existed and is present in the financial industry. However, the rise of technology and the use of online transactions has enhanced the presence of the impact of fraud in this industry. The increase in financial fraud cases in a technological era result from a lack of inter-organizational synergy and the privacy concerns that entail by making data available. On top of the increasing fraud cases, organizations are exposed to increasing regulatory, financial, reputational, and legal risks. Hence, the financial crime industry and fraud prevention organizations must act on this threat. Therefore, these actors need to improve their current workflow continuously to keep up with the new developments. Different studies propose that it is a potential opportunity to take the chance and bundle data together to learn from the existing environment and improve their workflows and prediction models. However, the main concern is that parties are reluctant to share data as it involves confidential and sensitive data, which malicious parties can leverage and abuse. Also, the increasing focus on privacy protection regulations makes it complex and challenging to exchange data easily. The dataset that actors are providing will contain personally identifiable information, which in fact cannot be shared and proposed without any legitimate reason and is subjected to the data privacy regulations.
The existing set of techniques for sharing and analysing data securely, such as differential privacy, homomorphic encryption and federated learning have been proposed in studies and use cases are built in the real world. However, these techniques are insufficient and capable enough to facilitate multi-actor (data owners) data exchange and analysis. Secure Multiparty Computation, however, is capable of having multiple data owners securely perform a joint analysis. For this reason, this study has been focused on SMPC. In the case of the financial crime industry, it requires the involvement of multiple stakeholders sharing data simultaneously. Especially in the case of banks, it is essential to have bundled data for all transactions as these are connected.
To understand the purpose and the concept of the study, it is essential to have a basic understanding of SMPC. Secure Multiparty Computation is a cryptographic method for parties to mutually compute a function over all parties' acquired inputs while keeping the intake private throughout the entire process. Independent computation nodes will perform and provide analysis outcomes to designated parties. The concept of Secure Multiparty Computation has been studied by academia since the 1980s (Yao, 1982). However, the applications and introduction of SMPC are relativity novel to organizations and will not be immediately accepted. A technique such as SMPC will require participating parties to share a mutual interest and willingness to contribute continuously. It is uncertain if organizations will accept and adopt SMPC as mentioned before. Therefore, the study will also incorporate the concept and theory of collective action to understand the motives and the common goal for stakeholders in the anti-fraud industry to accept the technique and collaborate. A common goal, also known as a collective goal or interest, would create acceptance among the group. In this case, it will help to identify the factors and interests that influence an organization's decision to engage in collective action for developing MPC for fraud detection in the financial industry...