4 ms·

>One recent scheme involves getting consumers to say “yes” and later using a recording of the response to allow unauthorized charges on the person’s credit card

by ikeboy 9y ago

>One recent scheme involves getting consumers to say “yes” and later using a recording of the response to allow unauthorized charges on the person’s credit card account, the F.C.C. warned in March. I doubt banks are using voice recognition on the single word "yes" to verify charges. Does this make any sense?

adventured 9y ago
That's a case where it has happened a small number of times, complaints were filed, the FCC & Co. picked up on the scam, and they view it as their responsibility to keep the public safe so they generalize. Although I suppose given the speed at which AI is improving, the roboscammers will be utilizing more and more advanced methods soon when it comes to voice sample manipulation. Most of this criminal activity is coming from outside the US, courtesy of Twilio & Co. having turned the phone number system into an API such that it's now easy to rapidly buy and ditch numbers for cheap. The only practical solution for the end-user point is to insert greater intelligence into the system at the carrier level or at the phone level. We're facing a problem where very large numbers of phone numbers will be black listed given enough time, unless we insert a greater cost into the system to make the spamming unprofitable (or make it much harder to get a phone number via services like Twilio). The US could try targeting the criminal organizations and individuals operating these scams on the other end; of course that's expensive and slow, and those people will just keep popping up so long as it's easy to do and profitable. The improving and cheapening AI will make it more and more profitable, as the scammers learn to snake into new, more lucrative financial outlets using authentications / identities they steal.
- icebraining 9y ago
  Why would they need Twillio et all to give them new numbers, if as everyone (including in this thread) says they can spoof the number, even showing as the receiver's number itself?

by ikeboy 9y ago

adventured 9y ago

That's a case where it has happened a small number of times, complaints were filed, the FCC & Co. picked up on the scam, and they view it as their responsibility to keep the public safe so they generalize. Although I suppose given the speed at which AI is improving, the roboscammers will be utilizing more and more advanced methods soon when it comes to voice sample manipulation. Most of this criminal activity is coming from outside the US, courtesy of Twilio & Co. having turned the phone number system into an API such that it's now easy to rapidly buy and ditch numbers for cheap. The only practical solution for the end-user point is to insert greater intelligence into the system at the carrier level or at the phone level. We're facing a problem where very large numbers of phone numbers will be black listed given enough time, unless we insert a greater cost into the system to make the spamming unprofitable (or make it much harder to get a phone number via services like Twilio). The US could try targeting the criminal organizations and individuals operating these scams on the other end; of course that's expensive and slow, and those people will just keep popping up so long as it's easy to do and profitable. The improving and cheapening AI will make it more and more profitable, as the scammers learn to snake into new, more lucrative financial outlets using authentications / identities they steal.

icebraining 9y ago
Why would they need Twillio et all to give them new numbers, if as everyone (including in this thread) says they can spoof the number, even showing as the receiver's number itself?