IBM develops audio-masking technology

IBM's India Research Laboratory (IRL) has developed technology that automatically detects and masks sensitive information in audio recordings.

The technology is expected to be useful for call centre operations which record conversations between call centre staff and customers for a number of reasons, including monitoring of service quality, said Guruduth Banavar, director of IRL, in a telephone interview on Wednesday. Some of these audio recordings are also used to train new staff, he added.


The technology utilizes a combination of speech analytics and metadata to locate and mask portions of an audio recording during playback to individuals that are not authorized to hear the sensitive information, according to Banavar. The information that is to be masked can be configured depending on the requirement, and the masked portions can be presented in many ways, such as white noise, silence or an announcement that the information has been edited, he added.

The ability to maintain customer trust requires organizations to be able to ensure the security of their customer's private information, such as credit card numbers, personal identification numbers (PIN), social security numbers and other information collected through interactions between call centre staff and customers, IBM said.

IBM is currently running pilots of the technology within the company. A decision on commercializing the technology will be taken by the business units in IBM, Banavar said. The technology has applications in a number of other areas, such as medical diagnosis, where recorded information collected in one context is later used for training people, he added.




PhoneValet 5.4 improves iPhone/Touch audio
LG ships iPod-ready home theater systems
Cut price international calls for iPhone users