Global banking cooperative SWIFT is poised to introduce advanced AI-driven fraud detection systems by 2025, aiming to enhance the security framework of financial institutions globally. This strategic initiative is designed to provide financial organisations with real-time, accurate insights into potential fraudulent activities, leveraging cutting-edge artificial intelligence to address a rising tide of sophisticated financial crimes.
In recent years, the financial sector has grappled with increasingly complex fraud schemes, including AI-generated deepfakes impersonating high-profile executives and synthetic identities crafted by artificial intelligence. These innovative criminal tactics have contributed to a staggering global financial fraud cost of $485 billion in 2023, as reported by industry estimates.
Jerome Piens, the Chief Product Officer at SWIFT, highlighted the urgent need for the financial industry to bolster its defence mechanisms against these threats. "Bad actors are using increasingly sophisticated tactics to commit financial crime," Piens stated, emphasising the necessity for enhanced protective measures to maintain global transaction confidence.
The forthcoming technology is expected to be a significant enhancement of SWIFT's Payment Control Service, already widely utilised by small and midsize financial institutions. SWIFT's AI engine will utilise the vast volumes of transactional data traversing its networks each year. This data will serve as a learning base for the AI system, with privacy safeguarded by anonymising personal identifiers and replacing them with pseudonyms. The AI technology aims to detect suspicious patterns in real time, thus mitigating fraud risk and ensuring a secure banking environment, according to John McHugh, head of CIB Operations and Control at Standard Bank.
Earlier in the year, banks across the Asia-Pacific, European, Middle Eastern, and North American regions participated in a successful pilot programme with SWIFT. The outcome of these pilots contributed to the final design of the tool, readying it for wider deployment.
Beyond the immediate implementation of this technology, SWIFT is exploring collaborative data-sharing methods with major financial entities. One promising area of research is federated learning, a machine learning approach that allows model training across various decentralised servers, ensuring that proprietary data remains secure.
Data sharing was also a focal topic at the 2024 Sibos conference in Beijing, where a panel discussion featuring representatives from SWIFT, Deutsche Bank, Intesa Sanpaolo, and UniCredit explored the potential benefits and challenges. A key takeaway was the call for regulatory bodies to establish a minimum data-sharing framework to facilitate fraud detection. Currently, the absence of such regulations leaves financial institutions cautious about sharing transaction data, despite the potential benefits in enhancing fraud detection capabilities.
As SWIFT prepares for the deployment of this AI-driven technology, it represents a significant step forward in the pursuit of a more secure global banking environment, setting the stage for future innovations in financial crime prevention.
Source: Noah Wire Services