Good morning, it's finance editor Jeff Roberts tagging in for Sheryl. It's been a year since clever crooks tricked a staff member of a design firm into joining a video call with deep fake versions of the company's CFO and other executives—a call that was so convincing that the employee wired a series of payments totaling over $25 million. The incident, which sounds like something ripped out of a Black Mirror episode, likely kept more than a few finance chiefs up at night. And for good reason.
On a recent webinar, the CEO of fintech firm Plaid said fraud has soared to the top of the list of things that banks are focusing on this year. The reason for this, unsurprisingly, is AI, which has given bad actors a powerful set of tools to trick finance departments into approving fraudulent payments. Those tools include deepfake videos like the one described above, but also ones to create hyperrealistic replicas of corporate documents. It's also easy enough to imagine hackers getting inside a CFO's email in order to train generative AI models to mimic their communication styles.
It's frightening stuff, but the news isn't all bad. I recently spoke with Mastercard CFO Sachin Mehra, who described the global payment giant's AI strategy—including when it comes to fraud. "Technology works both ways," he said. "Right now, if it helps us, it also helps the people who want to do bad acts and in this case, people who want to do fraudulent activity, cyber crime, or whatever the case might be."
Mehra's comment about new tech being a two-way street is correct and, when it comes to the current wave of fraud, it means that companies who stay on top of their AI game will be in a good position to defend themselves. In Mastercard's case, it helps that the company has been honing its custom AI stack for years, in part thanks to the 2017 acquisition of a San Francisco artificial intelligence startup.
Mehra also says Mastercard and its clients benefit from the enormous scale of the company's network, which allows it to detect new types of fraud attacks. "We get to see the data not only for a single customer, but we get to see it network-wide...We have trained our models in a manner to be able to, with a high degree of efficacy, identify when there is a network level fraud event which might be taking place."
The upshot is that, by combining its scale with sophisticated AI tools, Mastercard can serve as an early warning system for its customers—a feature that Mehra likened to a safety net. This underscores how, for Mastercard and other companies, the current wave of AI fraud is not just a massive threat but also a business opportunity for those who can help to stop it.
