![](\"https:\/\/static01.nyt.com\/images\/2023\/06\/30\/multimedia\/30bank-deepfakes-jcwf\/30bank-deepfakes-jcwf-articleLarge.jpg\")
<\/p>\nThis spring, Clive Kabatznik, an investor in Florida, called his local Bank of America representative to discuss a big money transfer he was planning to make. Then he called again.<\/p>\n
Except the second phone call wasn\u2019t from Mr. Kabatznik. Rather, a software program had artificially generated his voice and tried to trick the banker into moving the money elsewhere.<\/p>\n
Mr. Kabatznik and his banker were the targets of a cutting-edge scam attempt that has grabbed the attention of cybersecurity experts: the use of artificial intelligence to generate voice deepfakes, or vocal renditions that mimic real people\u2019s voices.<\/p>\n
The problem is still new enough that there is no comprehensive accounting of how often it happens. But one expert whose company, Pindrop, monitors the audio traffic for many of the largest U.S. banks said he had seen a jump in its prevalence this year \u2014 and in the sophistication of scammers\u2019 voice fraud attempts. Another large voice authentication vendor, Nuance, saw its first successful deepfake attack on a financial services client late last year.<\/p>\n
In Mr. Kabatznik\u2019s case, the fraud was detectable. But the speed of technological development, the falling costs of generative artificial intelligence programs and the wide availability of recordings of people\u2019s voices on the internet have created the perfect conditions for voice-related A.I. scams.<\/p>\n
Customer data like bank account details that have been stolen by hackers \u2014 and are widely available on underground markets \u2014 help scammers pull off these attacks. They become even easier with wealthy clients, whose public appearances, including speeches, are often widely available on the internet. Finding audio samples for everyday customers can also be as easy as conducting an online search \u2014 say, on social media apps like TikTok and Instagram \u2014 for the name of someone whose bank account information the scammers already have.<\/p>\n
<\/p>\n
\u201cThere\u2019s a lot of audio content out there,\u201d said Vijay Balasubramaniyan, the chief executive and a founder of Pindrop, which reviews automatic voice-verification systems for eight of the 10 largest U.S. lenders.<\/p>\n
Over the past decade, Pindrop has reviewed recordings of more than five billion calls coming into call centers run by the financial companies it serves. The centers handle products like bank accounts, credit cards and other services offered by big retail banks. All of the call centers receive calls from fraudsters, typically ranging from 1,000 to 10,000 a year. It\u2019s common for 20 calls to come in from fraudsters each week, Mr. Balasubramaniyan said.<\/p>\n
So far, fake voices created by computer programs account for only \u201ca handful\u201d of these calls, he said \u2014 and they\u2019ve begun to happen only within the past year.<\/p>\n
Most of the fake voice attacks that Pindrop has seen have come into credit card service call centers, where human representatives deal with customers needing help with their cards.<\/p>\n
Mr. Balasubramaniyan played a reporter an anonymized recording of one such call that took place in March. Although a very rudimentary example \u2014 the voice in this case sounds robotic, more like an e-reader than a person \u2014 the call illustrates how scams could occur as A.I. makes it easier to imitate human voices.<\/p>\n
A banker can be heard greeting the customer. Then the voice, similar to an automated one, says, \u201cMy card was declined.\u201d<\/p>\n
\u201cMay I ask whom I have the pleasure of speaking with?\u201d the banker replies.<\/p>\n
\u201cMy card was declined,\u201d the voice says again.<\/p>\n
The banker asks for the customer\u2019s name again. A silence ensues, during which the faint sound of keystrokes can be heard. According to Mr. Balasubramaniyan, the number of keystrokes correspond to the number of letters in the customer\u2019s name. The fraudster is typing words into a program that then reads them. <\/p>\n
In this instance, the caller\u2019s synthetic speech led the employee to transfer the call to a different department and flag it as potentially fraudulent, Mr. Balasubramaniyan said.<\/p>\n
Calls like the one he shared, which use type-to-text technology, are some of the easiest attacks to defend against: Call centers can use screening software to pick up technical clues that speech is machine-generated.<\/p>\n
\u201cSynthetic speech leaves artifacts behind, and a lot of anti-spoofing algorithms key off those artifacts,\u201d said Peter Soufleris, the chief executive of IngenID, a voice biometrics technology vendor.<\/p>\n
But, as with many security measures, it\u2019s an arms race between attackers and defenders \u2014 and one that has recently evolved. A scammer can now simply speak into a microphone or type in a prompt and have that speech very quickly translated into the target\u2019s voice.<\/p>\n
Mr. Balasubramaniyan noted that one generative A.I. system, Microsoft\u2019s VALL-E, could create a voice deepfake that said whatever a user wished using just three seconds of sampled audio.<\/p>\n
On \u201c60 Minutes\u201d in May, Rachel Tobac, a security consultant, used software to so convincingly clone the voice of Sharyn Alfonsi, one of the program\u2019s correspondents, that she fooled a \u201c60 Minutes\u201d employee into giving her Ms. Alfonsi\u2019s passport number.<\/p>\n
The attack took only five minutes to put together, said Ms. Tobac, the chief executive of SocialProof Security. The tool she used became available for purchase in January.<\/p>\n
While scary deepfake demos are a staple of security conferences, real-life attacks are still extremely rare, said Brett Beranek, the general manager of security and biometrics at Nuance, a voice technology vendor that Microsoft acquired in 2021. The only successful breach of a Nuance customer, in October, took the attacker more than a dozen attempts to pull off.<\/p>\n
Mr. Beranek\u2019s biggest concern is not attacks on call centers or automated systems, like the voice biometrics systems that many banks have deployed. He worries about the scams where a caller reaches an individual directly.<\/p>\n
\u201cI had a conversation just earlier this week with one of our customers,\u201d he said. \u201cThey were saying, hey, Brett, it\u2019s great that we have our contact center secured \u2014 but what if somebody just calls our C.E.O. directly on their cellphone and pretends to be somebody else?\u201d<\/p>\n
That\u2019s what happened in Mr. Kabatznik\u2019s case. According to the banker\u2019s description, he appeared to be trying to get her to transfer money to a new location, but the voice was repetitive, talking over her and using garbled phrases. The banker hung up.<\/p>\n
\u201cIt was like I was talking to her, but it made no sense,\u201d Mr. Kabatznik said she had told him. (A Bank of America spokesman declined to make the banker available for an interview.)<\/p>\n
After two more calls like that came through in quick succession, the banker reported the matter to Bank of America\u2019s security team, Mr. Kabatznik said. Concerned about the security of Mr. Kabatznik\u2019s account, she stopped responding to his calls and emails \u2014 even the ones that were coming from the real Mr. Kabatznik. It took about 10 days for the two of them to re-establish a connection, when Mr. Kabatznik arranged to visit her at her office.<\/p>\n
\u201cWe regularly train our team to identify and recognize scams and help our clients avoid them,\u201d said William Halldin, a Bank of America spokesman. He said he could not comment on specific customers or their experiences.<\/p>\n
Though the attacks are getting more sophisticated, they stem from a basic cybersecurity threat that has been around for decades: a data breach that reveals the personal information of bank customers. From 2020 to 2022, bits of personal data on more than 300 million people fell into the hands of hackers, leading to $8.8 billion in losses, according to the Federal Trade Commission.<\/p>\n
Once they\u2019ve harvested a batch of numbers, hackers sift through the information and match it to real people. Those who steal the information are almost never the same people who end up with it. Instead, the thieves put it up for sale. Specialists can use any one of a handful of easily accessible programs to spoof target customers\u2019 phone numbers \u2014 which is what likely happened in Mr. Kabatznik\u2019s case.<\/p>\n
Recordings of his voice are easy to find. On the internet there are videos of him speaking at a conference and participating in a fund-raiser.<\/p>\n
\u201cI think it\u2019s pretty scary,\u201d Mr. Kabatznik said. \u201cThe problem is, I don\u2019t know what you do about it. Do you just go underground and disappear?\u201d<\/p>\n