Most recently faced the news: «In Moscow, fraudsters make loans with the help of biometrics,» in which it was reported:
«In Moscow, multiple cases of using customer votes by fraudsters for loans or other financial products are recorded.
The attackers found a new way to deceive people in Moscow. They began to record the voices of bank customers, and after using them for loans. «
In general, this is not the first time when I meet reports that fraudsters collect voice samples for use in fraudulent schemes. Messages that calls from unknown numbers are coming, where the robot asks to answer questions about the answers «yes» or «no», and that these answers record fraudsters.
Let’s figure it out, is it possible to realize the financial operation using only voice?
Unified biometric system (EBS)
Those. the opportunity as such is there, but as they say
Collected and transmit to a single biometric system (EBS) biometric data only at the request and harmonization of the client, and if you have not been transmitted them initially, your voice does not help the fraudster.
In addition to the voice to identification, the photo is still used, while the special program reads the distance from the nose to the eye, from the nose to the lips, that each person is individually.
For remote identification, a combination of voice, face and some other parameters is used, while protection is used that excludes the photograph substitution.
It turns out only voice teams, such, for example, as «yes», «no» or «confirm», not enough to make banking operations.
So for what scammers sample voice?
It is possible to collect data for the future. Deepfake technology is currently developing, capable of submitting a person on photography and video images. If this technology develops enough to deceive the algorithms of protection against substitution, then fraudsters will really be able to perform operations on behalf of the victim.
As for the recording of the answers «Yes», «No», «I confirm» and so on. My opinion — fraudsters it is for nothing, it is much more important for them to just get a sample of your voice, and this can be done and by recording a conversation when conducting a harmless social survey.
At the same time, the technology of substitution of voice already exists and is actively developing. For example, the project
And if there is an opportunity to smey a whole speech, then I think to pronounce the victim’s voice «yes», for the labor program will not be.