A relatively new kind of identify fraud is wreaking havoc with completely fake personas—and Silicon Valley startups are helping banks fight it.
Synthetic fraud occurs when credit applicants create fake identities as “real” people with credit bureaus, take on debt, and then skip out on the repayment. By contrast, victims of traditional identity theft tend to find out that someone is using their personal information by seeing a loan they don’t recognize on their credit report or getting a call from a collections agency.
Because of how lenders and their systems are designed to combat identity fraud involving real identities, synthetic fraud represents a mounting problem. The Federal Trade Commission recently said that synthetic fraud is “the fastest growing” type of identity fraud in the United States.
“Synthetic fraud has a far lower profile in the public mind than identity theft because most people know when they or someone [else] is a victim of identity theft, but there aren’t any synthetic identity victims that can talk about it,” said Naftali Harris, CEO of SentiLink, a startup that specifically targets this type of fraud.
This week SentiLink raised $14 million in Series A funding from Andreessen Horowitz, Nyca Partners, Goldcrest, Felicis Ventures, Caffeinated Capital, Affirm CEO Max Levchin, and Plaid CEO Zach Perret. Both Harris and his cofounder Max Blumenfeld previously led the risk decisioning and risk operations organizations at Affirm.
San Francisco-based SentiLink essentially ingests credit application data, compares it with millions of other credit applications to see if it connects to known fraudsters, and returns risk scores to the lender in real time. The startup is building a universal identity bureau that can help identify synthetic identities.
Harris told Cheddar that SentiLink works with “a number of the top U.S. banks,” which collectively lose around $2 billion per year from synthetic fraud. He said while banks now approach SentiLink for help, a mere two years ago almost no one in financial services had even heard of the problem. Lenders were certain they had best-in-class identity theft prevention systems that would continue to protect them.
“This problem is brand new,” said David Sica, a principal investor at Nyca Partners, which invested in SentiLink. “Banks see synthetic fraudsters as real customers, so they don’t appear in the fraud numbers. It’s treated as a default when they look at it today, rather than as fraud.”
“If I made up a fake John Smith, I would get John Smith his own email address, sign John Smith up for social media accounts,” explained Harris. “And when somebody tries to ask me if I’m actually John Smith, I’d be able to verify the email, phone number, [and] answer questions about John Smith’s history… because I’m the one that created John Smith’s entire history.”
In the lending industry, a bank’s willingness to lend to a consumer is a function of every other lender’s willingness to lend to that same person. So if a single lending institution allows a made-up identity into the credit system, that identity will likely be cleared by other lenders.
Because of how the overall lending system works, Harris and his investors are betting that the synthetic fraud problem will grow significantly in the next year.
But fraud—especially when it involves completely fake personas—isn’t something that can be easily fought. Bad actors only become more sophisticated as the systems designed to thwart them improve.
Harris knows the uphill battle SentiLink faces: “It’s won’t be easy, but it is possible.”