US Patent 11069349 Privacy-preserving voice control of devices

The speech command issued to a voice activated/controlled system is anonymized so that biometric voice data of the speaker may not be received by the voice activated/controlled system. A spoken audio command is converted to text, which is then converted to a synthesized voice signal. The synthesized voice signal is then provided to the voice-activated/controlled device. The synthesized voice signal may be provided to the voice-activated device within a sound shield or enclosure so that the original speech command issued by the speaker may not be received by the voice-activated/controlled system. In this way, the speaker's actual voice and related data may be kept private and secure.