According to panellists at the Singapore Fintech Festival 2024, artificial intelligence (AI) will continue to be a hot topic among financial institutions as they prioritise efficiency gains from their digital transformation in 2025.
“There is still an enormous opportunity for financial institutions to use AI for operational efficiency, and some of them are low-hanging use cases. For example, they can use generative AI to power chatbots so that no matter what questions are asked, you’ll get a contextual response,” says Ari Sarker, president of Asia Pacific at Mastercard.
To demonstrate the real-world value of AI, Tal Cohen, president of Nasdaq, outlines how the stock exchange has already benefitted from the technology. “AI plays a role across our entire software engineering lifecycle. We’ve run maybe 30 proof-of-concepts and put some of those AI pilots in production, which has led to productivity gains of between 10% and 20%,” he says.
He adds: “We’ve also developed and launched an AI-driven order type that is approved by the US Securities and Exchange Commission.” Order types are programmed instructions traders use to tell exchanges how to handle their trades. Nasdaq’s research shows that by using AI to monitor market behaviour and make real-time adjustments to holding periods, the AI order type achieves a 20.3% increase in fill rates and an 11.4% reduction in mark-outs.
Beyond improving productivity, Ant International is using AI to deliver new services to customers. Its CEO, Yang Peng, gave the example of AI being used to answer customer queries on wealth management and guide customers to make better decisions on insurance products on its platform.
Additionally, Ant International is using AI to streamline and secure cross-border transactions for nearly 100 million SMEs across more than 200 markets. One example is an AI-powered foreign exchange (FX) model that predicts currency exchange rates hourly, helping to reduce merchants’ transaction costs.
See also: Navigating the workforce shift for a GenAI-powered future
AI challenges
Like other technologies, AI is a double-edged sword, presenting challenges such as balancing the accuracy of large language models (LLMs) with privacy concerns. “We need to build contextual LLMs at an industry level [to improve the accuracy of generative AI’s responses]. But how can we do that when financial institutions don’t want to collaborate or are afraid of sharing data, especially sensitive ones like customer information? So, we need to think of how we can build an LLM that uses anonymised data [to address the issue of privacy] while still giving fantastic output and allowing companies to retain competitiveness,” says Sarker.
AI ethics, or using AI responsibly, is another key area that financial institutions should focus on when they embrace the technology.
See also: Updated version of Singapore's unified QR code, SGQR+, launched by NETS
Yang says that Ant Financial decided to consolidate 300 AI models into one transformer model, which is designed to comprehend context and meaning by analysing the relationship between different elements and is used in LLMs. “Although the move improved the LLM’s accuracy, it raises critical questions if we rely solely on that LLM. For instance, how do we safeguard our operations if something goes wrong with the LLM? Where should human intervention step in to address discrepancies in the model’s output? What policies should govern the use of multiple LLMs as backups, and how do we ensure a seamless transition from the old to the new LLM when necessary?”
Cohen adds that Nasdaq has made five key investments to harness the power of AI safely and responsibly. These include effective data management, a clear understanding of what data it holds and where it is located, embracing the cloud and adopting a mature approach to information security.
“We’ve also ensured good governance. Although many people perceive governance as an inhibitor, it’s actually a lubricant to innovation. Finally and most importantly, we’ve focused on upskilling our employees because we need them to go on [and support] the organisation’s AI journey,” he continues.
Stay ahead of threats
The panellists also believe cybersecurity will remain a top priority for financial institutions next year. Every company will face a cyber incident at some point, and its response will directly affect the trust of regulators and customers.
Yang says financial institutions must first prioritise cybersecurity and compliance to secure a seat at the boardroom table. Regular reviews of cybersecurity policies and procedures, he adds, are crucial to staying ahead of emerging cyber threats and ensuring alignment with evolving regulations, especially for firms operating across multiple markets.
For example, Ant International has developed an anti-deepfake electronic Know Your Customer (e-KYC) tool in response to the growing challenge of detecting deepfakes. As these synthetic media become increasingly difficult to identify, the tool has achieved an interception success rate of over 99%.
To stay ahead of the latest tech trends, click here for DigitalEdge Section
Additionally, financial institutions should hold regular tabletop exercises — discussion-based sessions designed to prepare key team members for emergencies, disasters or crises.
“At Nasdaq, we conduct tabletop exercises at all levels because everyone is accountable for cyber incidents. We also [educate everyone] about cyber threats, including the risks emerging technologies like AI bring so that we’re ahead of those threats,” says Cohen.
No single entity can effectively combat cyber threats alone, adds Sarker. Therefore, financial institutions must collaborate with regulators and law enforcement agencies to tackle cyber crimes like fraud and scams effectively.
