Why Speaker Verification?
As worldwide telecommunications markets have diversified and deregulated over the last 15 or so years, many important trends have emerged. Two such trends which have arguably caught the attention of the individual more than most, are the following; that nowadays, the consumer need not be tied to a particular operator for a particular long-distance call, and that it is now possible to access a wide-range of services at any time throughout the day or night, solely by telephone.
In the 90s, millions of people exercise this freedom of choice and take advantage of the convenience of the telephone, every time they use a telephone calling-card to place a long-distance call, or check their account balance with a telephone banking service. However, the introduction of these new possibilities has raised new questions; how can the consumer be sure that no-one else can gain access to his or her telephone calling card, or to the private details of his or her bank account?
Unfortunately, at the moment, the consumer can not be sure at all. A vast level of telephone calling card fraud is reported every year in the United States (the estimated figure for 1995 was $1 billion, with 1 in 6 calling card users estimated to be a victim of calling card fraud throughout the year), hitting both the telecommunications companies and the consumer, who is typically held responsible for the first $50 of unapproved calling-card charges. Fraudsters routinely loiter around public telephones, particularly the large banks of telephones found at railway stations, airports or hotels, simply observing the account numbers and personal identification number (PIN) codes dialled by users of calling-cards. Stolen codes can be sold on or used to set up "call-sell operation", where fraudsters operate at payphones and offer to set up telephone calls for passers-by for a flat fee.
The situation in direct banking is somewhat different. As yet, no direct banking-based frauds have come to light. This is because currently-available services are either relatively unattractive to fraudsters. Automatic services, in which a customer identifies himself to the system usually by typing his account number and a PIN code, typically limit themselves to offering details of the account holder's balance, a number of recent transactions, and at most allow the caller to move money between a checking account and a savings account.
Agent-assisted services, in which the caller speaks to an employee of the bank throughout the call, only verify the caller on the basis of a lengthy "security handshake" in which the caller answers personal questions to which only the genuine caller should know the answer. In addition to this procedure, major transactions are always delayed until they can be confirmed in writing.
Speaker verification (SV), the authorisation of a caller's identity claim using information derived from the sound of his or her voice, can be applied in both cases. It has the potential to improve security in telephone calling-card services, so that it is no longer enough for the fraudster to "shoulder-surf" an unwitting cardholder's details; the voice must be exactly right as well. Moreover, by using the appropriate technologies, not even a tape of the true cardholder's voice speaking these details is enough!
Even an imperfect SV system, which may in fact let in 5% or so of impostors, could make the call-selling business unattractive enough to deter most call-selling fraudsters. In the area of direct banking, SV can improve speed and friendliness of authentication in agent-assisted services and remove the need for transaction delay; in automatic services, it can allow a greater and more useful range of services, such as bill-payment, to be accessed simply and conveniently.
As part of its work in investigating speaker verification, the CAVE project has carried out a major survey of the market for speaker verification in direct banking and calling card services. You can download a public version of this report by clicking here.
