To perform the tasks requested by users in the digital economy, voice-based technologies record and process large amounts of data, including personal data, which is centrally stored in the cloud.
Voice-enabled systems capture the user’s speech signal (voice), but also background sounds!
Background sounds can provide private information regarding the user (e.g., the music he/she is listening to). Background conversations may also sometimes be recorded.
Users may be sharing personal information in their spoken messages, including words or utterances that explicitly reveal their name, address, preferences, etc., and more critical information such as credit card numbers.
When interacting with voice-enabled systems, users may also be revealing sensitive health-related data, for example when using voice-enabled apps to buy medical products such as pregnancy tests or simply by using e-health apps!
The user’s general traits (e.g., age, ethnicity) and states (e.g., level of intoxication, sincerity) can be inferred from voice characteristics to some extent.
It is even possible to infer the user’s identity from their biometric voice characteristics!
Voice-enabled apps may apply profiling techniques to deliver personalised responses to the users. Profiling works by providing customised answers based on the user’s preferences and past interactions.
Beware of spoofing! An artificial intelligence-generated voice deepfake has already been used to trick a CEO into transferring money from the company bank account to a fraudulent supplier.
According to a PWC survey, half of the people refusing to use voice technologies cite the need for privacy as a primary reason. Concerns include: (i) being continuously listened to and (ii) compromising their personal data.