Artificial intelligence is revolutionizing the finance industry, from operations to risk management to fraud detection. But when it comes to AI-driven decision-making, industry players need to ensure fairness and transparency.
That was a key message from panelists during a session on finance and AI. “I would say we’re in the trust business more than anything else,” said Darrell MacMullin, SVP of products and platforms with Mastercard Canada.
“It is important that we still treat AI as a guide, not as a god,” he said. “It is very, very important to establish the difference between those two things.”
Mastercard has embraced AI from the get-go. “We’ve invested in over 100 AI companies, we’ve acquired about 35, 36 of them, because you simply can’t build the DNA just within your own company that quickly,” said MacMullin.
“We don’t treat it as a product though,” he added. Instead, it’s treated as a ‘layer’ that’s woven into a number of different capabilities. For example, AI is being used to automate management of identities, accounts, devices and transactions—thanks to a rich data set (from about 125 billion payment transactions annually).
“From a fraud mechanism standpoint, that’s super important,” said MacMullin. “But the other end of that spectrum is around hyper-personalization and loyalty and engagement, and how do you optimize the experience for financial institutions, for consumers, for merchants, and even for governments?”
In order for companies to have a strong AI strategy, they need a comprehensive data strategy. “AI without data is meaningless,” said Petar Kramaric, CTO and co-founder of Flybits. That’s especially true in the finance space, where “regulation is everywhere.”
Purpose-built for financial services companies, Flybits keeps customer data secure and uses it responsibly—on the customer’s behalf—providing context-relevant information and engagements.
“With the advent of open banking, the end consumer is controlling and owning more of their own data,” said Kramaric. Consumer data is ‘currency,’ which means consumers want to know what they’re going to receive—from a product or services perspectives—in exchange for their data.
While data is important, so is transparency.
“Our analyst teams can cover 300 businesses today. Soon we’ll be able to cover thousands of businesses [with AI],” said Jean-Francois Bérubé, VP of quantitative strategies and data science with global investment group CDPQ.
“But transparency is extremely important. For the portfolio manager, it’s a must,” he said. “They must understand how decisions are made—what are the reasons why [AI] would recommend something.”
Portfolio managers—and anyone else working in finance—will need to be able to explain how they derived their conclusions, so having a ‘human-in-the-loop’ will be key.
Humans will still be “accountable in terms of the assets they’re dealing with,” said Bérubé. “Even though they have machines, machines will be influenced by human beings.”
