System and method for the detection and termination of fraudulent services

A fraud detection system for detecting fraudulent activity on a transaction network, for which each transaction over the network has an identifier associated therewith, the system comprising:

• a voice comparator for comparing a first sampled voice of a user of one of the identifiers with a subsequently sampled voice of a user of the same identifier taken during a subsequent use of said same identifier;
• a profiler for building a profile of usage by users of said same identifier based on the voice comparisons, such that the profile is representative of a number of different users of the identifier;
• and,
• a fraud detector for comparing the number of different users represented in the profile as using the identifier to a threshold indicative of fraudulent use, wherein in the event that the number of users exceeds the threshold a fraud alert signal is generated.

A method of automatic detection of fraudulent activity on a transaction network, for which each transaction over the network has an identifier associated therewith, the method comprising:

• sampling a first voice of a user of one of the identifiers;
• subsequently sampling a second voice of a user of the same identifier;
• comparing the first sampled voice with the subsequently sampled voice;
• profiling usage by users of said same identifier based on the voice comparisons, such that the profile is representative of a number of different users of the identifier;
• comparing the number of different users represented in the profile as using the identifier to a threshold indicative of fraudulent use;
• and,
• generating a fraud alert signal in the event that the number of users exceeds said threshold.

A fraud detection system for detecting fraudulent activity on a transaction network, for which each transaction over the network has an identifier associated therewith, the system comprising:

• a use characteristics fraud detection engine for analysing characteristics of the use of the network to determine whether fraud is indicated in the use of one of the identifiers;
• a voice based fraud detection engine for sampling the voice of each user of the identifier during each use of the identifier using a voice analysis technique to identify the number of unique users of the identifier and to indicate the presence of fraudulent use when the number of unique users of the identifier exceeds a threshold;
• and,
• a fraud alert generator to signal presence of fraudulent activity when both of the use characteristics fraud detection engine and the voice based fraud detection engine indicate the presence of fraud.

A method of automatic detection of fraudulent activity on a transaction network, for which each transaction over the network has an identifier associated therewith, the method comprising:

• analysing characteristics of the use of the network to determine whether fraud is indicated in the use of one of the identifiers;
• sampling the voice of each user during each use of said identifier and using a voice analysis technique to identify the number of unique users of the identifier to indicate the presence of fraudulent use in the event that the number of unique users of the identifier exceeds a threshold indicative of fraudulent use;
• and,
• generating a fraud alert for signalling presence of fraudulent activity when both of the use characteristics analysis and the voice based analysis indicate the presence of fraud.