Speech Technology Magazine

 

Banking on Voice ID

This is the first of a series of columns that Judith Markowitz will be writing on voice ID applications, markets, technology and products.
By Judith Markowitz - Posted Jun 30, 1997
Page1 of 1
Bookmark and Share

Voice ID is a generic term covering methods and technologies that use a person's voice for identification. The most familiar forms of voice ID are speaker verification and speaker identification. Speaker verification determines whether a person is who he or she claims to be. Speaker identification assigns an identity to the voice of an unknown person.


The impetus for creating a regular column that deals specifically with such technologies is that concern about personal property and information security has permeated our lives. We are not surprised by reports that losses from telephone toll fraud have reached $3 million a year; because even if we have not experienced cloning of a cellular telephone, we know someone who has. Nor are we shocked to learn that more than two-thirds of organizations surveyed last year by Ernst & Young about data security lost money from security breaches, or that MasterCard has been testing fingerprinting on credit cards. After all, credit card fraud and unauthorized access to sensitive data on computer networks have become distressingly familiar.

It is especially appropriate that this column is beginning in the current issue of Speech Technology because of the editorial focus on the financial services industry. Given its heavy reliance on computer storage, electronic transactions and telephone banking, the financial services industry worldwide feels particularly vulnerable to breaches of security. Accompanying the industry's move to alternative customer and inter-bank delivery channels is its examination of security strategies better suited to the new environments and more able to resist sophisticated attack.

The financial services industry's exploration of new forms of security has generated interest in biometrics. Speaker verification, in particular, is seen as a solution that could be applied across channels. Among the most widely discussed examinations of speaker verification are the pilot tests by Chase Manhattan Bank and the product testing done by Citicorp.

Chase did marketing research on customer acceptance of several biometric technologies, followed by two tests of speaker verification with more than 9,000 customers. (See the article in this issue by Elizabeth Boyle.)

Citicorp contracted Sandia National Laboratories to test several speaker verification systems over long-distance telephone channels. At the same time, Caller Verification in Banking and Telecommunications (CAVE) was beginning its operations in Europe. CAVE is a two-year project funded by the European Union to evaluate speaker verification in real-world banking and telecommunications applications.

This year certainly promises to be an exciting one for voice ID in financial services. Several departments within Citicorp are reportedly moving ahead with plans to implement speaker verification, CAVE has started to deploy its demo applications, and brokerage houses are beginning to examine speaker verification. What better reasons could there be for starting a column and newsletter than these?



Judith Markowitz, speech and speaker recognition industry expert and author of Voice ID: Applications and Resources for Speaker Recognition, is now publishing Voice ID Quarterly and is the co-author of Voice ID: Source Profiles and Voice ID in the Financial Services Industry.

Page1 of 1