Speaking like a high-pitched child can fool voice recognition systems

A new study finds that high-pitched voices work best for deceiving voice recognition systems.
  • A new study from Juniper Research has found that smart speakers such as Amazon Echo and Google Home will be installed in over 70 million US households by 2022, reaching 55% of all homes.
  • A number of banks have adopted voice recognition to fight fraud by using voice prints: unique biometric markers that can be made up of more than 100 characteristics, based on the physical configuration of the speaker’s mouth and throat.
  • However, identical twins have duped the system in the past and researchers from the University of Eastern Finland have now found that faking a squeaky kid’s voice or lowering your tone like an elderly person can also fool systems.
  • The research team analyzed speech from two professional impersonators who mimicked eight Finnish public figures. They say impersonators were able to bypass state-of-the-art voice recognition systems.
  • They also studied acted speech from 60 Finish speakers impersonating children and elderly people and found that high-pitched voices worked best for deceiving the systems.
  • The Finnish team note the possibility of other voice attacks, including voice conversion, speech synthesis and replay attacks. The research community is “systematically developing techniques and countermeasures against technically generated attacks,” they say.

So you Think Shes Cheating on You - Put Your Mind at Rest With FlexiSPY (728x90)

Be the first to comment

Leave a Reply

Your email address will not be published.