Privacy campaigners are alarmed at the rise and spread of voiceprint technology that can accurately identify individuals from their voice alone.
According to a survey by the Associated Press of voice biometrics, more than 65 million voiceprints are already stored on databases worldwide, and will be the ‘de facto standard in the next two or three years’. Barclays bank, for example, has been experimenting with using voiceprint IDs for its wealthiest customers, and plans to extend the ID to its 12 million retail banking customers,
‘The general feeling is that voice biometrics will be the de facto standard in the next two or three years.’
Several governments have started to collect voice biometrics, most notably Turkey, where cell phone provider Turkcell has stored 10 million voiceprints. South Africa’s Social Security Agency with 7 million voiceprints, and the New Zealand Internal Revenue Department with 1 million voiceprints, are catching up however.
It is also proving a hit with companies which wish to take advantage of the technology for marketing purposes. As Professor Joseph Turow, a privacy and surveillance expert at the University of Pennsylvania, explained to the Guardian,
‘Companies are using data drawn from our internet and purchasing behaviour – and now our voices – and connecting it to the identities that they’ve created for us. Then they can lead us in a variety of different directions, based on their stereotype of us.’
This rise in the use of voiceprint technology, which is now estimated to generate $400 million, and is expected to rise to between $730m and $900m next year, was confirmed by Paul Burmester, CEO of ValidSoft,
‘There’s a misconception that the technology we have today is only in the domain of the intelligence services, or the domain of Star Trek. The technology is here today, well-proven and commonly available.’
Jay Stanley, an expert on technology-related privacy issues at the American Civil Liberties Union expressed concern at this huge and largely undocumented rise the use of a technology that has massive privacy implications,
‘This suggests there is a major new biometric tool that is being rolled out with very little public discussion. Obviously fraud protection is a good thing, but it raises implications that need to be looked into.’
Electronic Frontier Foundation staff attorney Lee Tien explained that ‘there is already discussion’ on using voice sensors in public places which could identify individuals and triangulate results to discover their physical location with a very high degree of accuracy,
‘Even where the technology wasn’t designed for eavesdropping or tracking people, it could still identify them and associate them with a location.’
Further fears have been expressed over how users of voice services that rely on anonymity, such as suicide hotlines and domestic violence counselling services, can know their identities will not be compromised, while Stanley noted that inaccuracies in the technology could result in unfortunate unintended consequences,
‘Biometrics are never 100% accurate. Are people going to be blacklisted by government institutions because their voice is mistaken for that of a fraudster?’
Irish privacy researcher Sadhbh McCarthy, painted a very grim picture of the future if voice biometric collection continues unabated,
‘It’s more mass surveillance. The next thing you know, that will be given to border guards, and you’ll need to speak into a microphone when you get back from vacation.’