Systems and methods for detecting suspicious voice calls
The disclosed computer-implemented method for detecting suspicious voice calls may include (i) identifying an incoming voice call, (ii) extracting, from audio of the incoming voice call, a plurality of characteristics, (iii) calculating a trustworthiness score of the plurality of the characteristics based on a response by a recipient of the incoming voice call, and (iv) storing the trustworthiness score of the plurality of characteristics in a reputation database that (a) receives a request for the trustworthiness score, the request originating from an additional computing device and including an additional plurality of characteristics extracted from an additional incoming voice call, (b) determines that the additional plurality of characteristics matches the plurality of characteristics, and (c) enables the additional computing device to perform a security action on the additional incoming voice call by sending the trustworthiness to the additional computing device. Various other methods, systems, and computer-readable media are also disclosed.
1. A computer-implemented method for detecting suspicious voice calls, at least a portion of the method being performed by a computing device comprising at least one processor, the method comprising:
identifying, by the computing device, an incoming voice call;
extracting, by the computing device and from audio of the incoming voice call, a plurality of characteristics of the audio of the incoming voice call;
calculating a trustworthiness score of the plurality of the characteristics based at least in part on analyzing a recipient response to the incoming voice call that was made, within the incoming voice call, by a recipient of the incoming voice call and further at least in part on individual trustworthiness scores for each of a geolocation of the call, background noise of the call, a tone of voice of a caller of the call, and keywords used by the caller; and
providing the plurality of characteristics and the trustworthiness score of the plurality of characteristics to a reputation database that:
stores the plurality of characteristics and the trustworthiness score of the plurality of characteristics;
receives a request to evaluate an additional incoming voice call, wherein the request originates from an additional computing device and includes an additional plurality of characteristics extracted from audio of the additional incoming voice call to the additional computing device;
determines that the additional plurality of characteristics matches the plurality of characteristics; and
in response to determining that the additional plurality of characteristics matches the plurality of characteristics, enables the additional computing device to perform a security action on the additional incoming voice call by sending the trustworthiness score for the plurality of characteristics to the additional computing device.
2. The computer-implemented method of claim 1 , wherein:
the recipient response by the recipient of the incoming voice call comprises a negative response; and
calculating the trustworthiness score of the plurality of characteristics comprises reducing a trustworthiness score of the plurality of characteristics based on the negative response.
3. The computer-implemented method of claim 1 , wherein:
the recipient response by the recipient of the incoming voice call comprises a positive response; and
calculating the trustworthiness score of the plurality of characteristics comprises increasing a trustworthiness score of the plurality of characteristics based on the positive response.
4. The computer-implemented method of claim 1 , wherein extracting, by the computing device the plurality of characteristics of the audio of the incoming voice call comprises extracting at least one non-audio characteristic of the incoming call.
5. The computer-implemented method of claim 1 , wherein calculating the trustworthiness score of the plurality of characteristics comprises calculating an individual trustworthiness score for each characteristic in the plurality of characteristics.
6. The computer-implemented method of claim 1 , calculating the trustworthiness score of the plurality of characteristics comprises calculating a cluster trustworthiness score for the plurality of characteristics as a cluster.
7. The computer-implemented method of claim 1 , calculating the trustworthiness score of the plurality of characteristics comprises using a machine learning algorithm to calculate the trustworthiness score for the plurality of characteristics.
8. The computer-implemented method of claim 1 , wherein the recipient response from the recipient of the incoming voice call comprises the recipient of the incoming voice call terminating the incoming voice call within a predetermined time period of the start of the incoming voice call.
9. The computer-implemented method of claim 1 , wherein the recipient response from the recipient of the incoming voice call comprises a tone of voice of the recipient of the incoming voice call.
10. The computer-implemented method of claim 1 , wherein the recipient response from the recipient of the incoming voice call comprises at least one keyword used by the recipient of the incoming voice call.
11. A computer-implemented method for detecting suspicious voice calls, at least a portion of the method being performed by a computing device comprising at least one processor, the method comprising:
identifying, by the computing device, an incoming voice call;
extracting, by the computing device, from audio of the incoming voice call, a plurality of characteristics of the audio of the incoming voice call;
retrieving, from a reputation database, a trustworthiness score of the plurality of characteristics, wherein the reputation database:
identifies audio from a previously received voice call;
extracts an additional plurality of characteristics from a recording of audio of the previously received voice call
calculates an additional trustworthiness score of the additional plurality of characteristics based at least in part on analyzing a recipient response to the previously received voice call that was made, within the previously received voice call, by a recipient of the previously received voice call and further at least in part on individual trustworthiness scores for each of a geolocation of the call, background noise of the call, a tone of voice of a caller of the call, and keywords used by the caller; and
in response to determining that the plurality of characteristics matches the additional plurality of characteristics, fulfilling the request for the trustworthiness score of the plurality of characteristics by sending the additional trustworthiness score to the computing device;
determining, based at least in part on the trustworthiness score of the plurality of characteristics, that the incoming voice call is suspicious; and
performing, by the computing device, a security action on the incoming voice call in response to determining that the incoming voice call is suspicious.
12. The computer-implemented method of claim 11 , wherein performing the security action comprises terminating the incoming voice call.
13. The computer-implemented method of claim 11 , wherein performing the security action comprises warning a recipient of the incoming voice call that the incoming voice call is suspicious.
14. The computer-implemented method of claim 11 , wherein determining, based at least in part on the trustworthiness score of the plurality of characteristics, that the incoming voice call is suspicious comprises:
calculating a total trustworthiness score by summing an individual trustworthiness score of each characteristic in the plurality of characteristics; and
determining that the total trustworthiness score falls below a predetermined threshold for trustworthiness.
15. A system for detecting suspicious voice calls, the system comprising:
an identification module, stored in memory, that identifies, by a computing device, an incoming voice call;
an extraction module, stored in memory, that extracts, by the computing device and from audio of the incoming voice call, a plurality of characteristics of the audio of the incoming voice call;
a calculation module, stored in memory, that calculates a trustworthiness score of the plurality of the characteristics based at least in part on analyzing a recipient response to the incoming voice call that was made, within the incoming voice call, by a recipient of the incoming voice call and further at least in part on individual trustworthiness scores for each of a geolocation of the call, background noise of the call, a tone of voice of a caller of the call, and keywords used by the caller;
a storing module, stored in memory, that provides the plurality of characteristics and the trustworthiness score of the plurality of characteristics to a reputation database that:
stores the plurality of characteristics and the trustworthiness score of the plurality of characteristics;
receives a request to evaluate an additional incoming voice call, wherein the request originates from an additional computing device and includes an additional plurality of characteristics extracted from audio of the additional incoming voice call to the additional computing device;
determines that the additional plurality of characteristics matches the plurality of characteristics; and
in response to determining that the additional plurality of characteristics matches the plurality of characteristics, enables the additional computing device to perform a security action on the additional incoming voice call by sending the trustworthiness score for the plurality of characteristics to the additional computing device; and
at least one physical processor that executes the identification module, the extraction module, the calculation module, and the storing module.
16. The system of claim 15 :
wherein:
the identification module, identifies, by the computing device, an additional incoming voice call; and
the extraction module extracts, by the computing device, from audio of the additional incoming voice call, an additional plurality of characteristics of the audio of the additional incoming voice call;
further comprising:
a retrieving module, stored in memory, that retrieves, from the reputation database, a trustworthiness score of the additional plurality of characteristics;
a determination module, stored in memory, that determines, based at least in part on the trustworthiness score of the additional plurality of characteristics, that the additional incoming voice call is suspicious; and
a security module, stored in memory, that performs, by the computing device, a security action on the additional incoming voice call in response to determining that the incoming voice call is suspicious.
17. The system of claim 16 , wherein the security module performs the security action by terminating the incoming voice call.
18. The system of claim 15 , wherein the calculation module calculates the trustworthiness score of the plurality of characteristics by calculating an individual trustworthiness score for each characteristic in the plurality of characteristics.
19. The system of claim 15 , wherein the calculation module calculates the trustworthiness score of the plurality of characteristics by calculating a cluster trustworthiness score for the plurality of characteristics as a cluster.
20. The system of claim 15 , wherein the calculation module calculates the trustworthiness score of the plurality of characteristics by using a machine learning algorithm to calculate the trustworthiness score for the plurality of characteristics.