As rapid advancements in generative artificial intelligence (AI) continue to take the global financial sector by storm, questions about how to leverage these technologies in a responsible manner are becoming as important as their potential to unlock new benefits in banking and payments. In a recent panel held at Volante’s 2023 EVOLVE conference, we delved directly into this very topic, exploring the unique ethical challenges to implementing AI-based solutions.
What is meant by “responsible” AI in banking & payments?
Generative AI is already revealing its transformative potential in many ways throughout the financial sector, particularly when it comes to eliminating manual processes and boosting productivity. However, the ability to retrieve and analyze documents in a matter of seconds as opposed to minutes or even hours may come at a cost if we don’t take measures to ensure these tools aren’t used irresponsibly.
When asked what would constitute an irresponsible use of generative AI in the context of banking, panelists agreed that the primary concern was around the potential for AI to produce biased or inaccurate information. More specifically, many of these systems are being trained to solve specific problems, but that doesn’t mean that what an AI deems to be the most effective solution won’t also yield negative results.
“Imagine if we want to train AI on smart routing and to find the most cost-effective and efficient way to send a payment, but then the Gen-AI suggests routing payments through North Korea,” remarked Chris Nichols, Director of Capital Markets, SouthState Bank. “I mean that could be an absolute disaster for our financial ecosystems.”
Given that the knowledge of AI systems today is based on data being fed in by users, the panelists further agreed that what makes an AI responsible or irresponsible ultimately comes down to how it’s being trained. “It’s almost a bit like parenting,” noted Alan Ng, Managing Director Payments, Accenture. “AI is like a kid, and how the kid turns out really depends on how they’re raised and the information they’re being fed. It could either turn out to be a good kid or a monster, so I think that’s the challenge we are facing and we are at that inflection point right now.”
What kind of regulations or internal actions are needed to ensure the responsible use of AI?
Understanding that the use of generative AI systems in banking and payments could result in harm to both institutions and their clients, whether due to the proliferation of misinformation or unchecked biases, the question remains: whose responsibility is it to ensure these technologies are used in a positive, ethical manner?
Despite the typically embattled relationship between financial institutions and regulators, all of our panelists agreed on the need for some level of regulatory oversight to support the advancement of AI-based technologies. “The government involvement has to be there,” said Ng, comparing future AI regulations to those implemented in adjacent areas such as social media and blockchain. “This might temper some of the innovations in the space, but I think this is a natural evolution and one we’re seeing with many other technologies.”
Importantly, however, no one believed that the responsibility for ethical AI use rests solely on government regulations, but rather on a multi-tiered approach that involves basic self-governance, as well as going above and beyond to ask hard questions about how this technology can be used specifically to create positive outcomes.
“It’s not enough to be unbiased [regarding AI], the question is how can we use this technology for good,” said Aniruddh Singh, Executive Director, Wholesale Architecture, Nomura Holding America. “I think this is causing a healthy discussion between regulators and businesses about how we turn generative AI into a positive force and what that means. And I think this is an ethical question that needs to be answered.”
To learn more about Volante Evolve 2023, please visit our event web page.