Routing of sensitive-information utterances through secure channels in interactive voice sessions

1. A computer-implemented method comprising:

obtaining, by one or more processors of a first computing system, first audio data representing one or more initial utterances during an interactive voice session with an interactive voice system that interacts with a server to obtain content to provide to a user during the interactive voice session, wherein the interactive voice system is configured to execute on a second computing system and the server is configured to execute on a third computing system;

transmitting, by the one or more processors of the first computing system, the first audio data to the interactive voice system via a first communication channel;

generating, by the one or more processors of the first computing system, based on the first audio data, a prediction that a second audio data representing a subsequent utterance of the user in the interactive voice session will contain sensitive information, wherein the subsequent utterance follows the one or more initial utterances in time;

upon obtaining, by the one or more processors of the first computing system, the second audio data that includes the sensitive information:

transmitting, by the one or more processors of the first computing system, third audio data to the interactive voice system via the first communication channel in place of the second audio data, wherein the third audio data represents data that replaces the sensitive information; and

transmitting, by the one or more processors of the first computing system, the second audio data to the server via a second communication channel that bypasses the interactive voice system.

2. The computer-implemented method of claim 1 , wherein:

the interactive voice system is configured to provide semantic content of the third audio data to the server, and

the server is configured to determine that the semantic content of the third audio data corresponds to the second audio data and perform an action based on the second audio data instead of the third audio data.

3. The computer-implemented method of claim 2 , further comprising:

sending, by the one or more processors of the first computing system, the third audio data to the server via the second communication channel, and causing the server to (1) generate a mapping of the second audio data and the semantic content of the third audio data, and (2) determine, based on the mapping, that the semantic content of the third audio data corresponds to the second audio data.

4. The computer-implemented method of claim 1 , wherein the third audio data is based on a voice of the user.

5. The computer-implemented method of claim 1 , further comprising:

obtaining, by the one or more processors of the first computing system, a timed volatile random subset (t-VRS) for the interactive voice session from the server; and

generating, by the one or more processors of the first computing system, the third audio data based on words in the t-VRS for the interactive voice session.

6. A computing system comprising:

memory; and

one or more processors implemented in circuitry and communicatively coupled to the memory, the one or more processors configured to:

obtain first audio data representing one or more initial utterances during an interactive voice session with an interactive voice system and that interacts with a server to obtain content to provide to a user during the interactive voice session, wherein the computing system is a first computing system, the interactive voice system is configured to execute on a second computing system and the server is configured to execute on a third computing system;

transmit the first audio data to the interactive voice system via a first communication channel;

generate, based on the first audio data, a prediction that a second audio data representing a subsequent utterance of the user in the interactive voice session will contain sensitive information, wherein the subsequent utterance follows the one or more initial utterances in time;

upon obtaining the second audio data that includes the sensitive information:

transmit third audio data to the interactive voice system via the first communication channel in place of the second audio data, wherein the third audio represents data that replaces the sensitive information; and

transmit, the second audio data to the server via a second communication channel that bypasses the interactive voice system.

7. The computing system of claim 6 , wherein the server comprises an application configured to operate on the third computing system.

8. The computing system of claim 6 , wherein:

the interactive voice system is configured to provide semantic content of the third audio data, and

the server is configured to determine that the semantic content of the third audio data corresponds to the second audio data and perform an action based on the second audio data instead of the third audio data.

9. The computing system of claim 8 , wherein:

the one or more processors are further configured to send the third audio data to the server via the second communication channel, causing the server to (1) generate a mapping of the second audio data and the semantic content of the third audio data, and (2) determine, based on the mapping, that the semantic content of the third audio data corresponds to the second audio data.

10. The computing system of claim 6 , wherein the third audio data is based on a voice of the user.

11. The computing system of claim 6 , wherein the interactive voice system is a voice assistant system.

12. The computing system of claim 6 , wherein the one or more processors of the first computing system are further configured to:

obtain a timed volatile random subset (t-VRS) for the interactive voice session from the server; and

generate the third audio data based on words in the t-VRS for the interactive voice session.

13. The computing system of claim 12 , wherein the one or more processors are further configured to receive refreshed t-VRS's from the server at configured intervals or in response to an event.

14. One or more non-transitory computer-readable storage media including instructions that, when executed by the one or more processors, cause the one or more processors of a first computing system to:

obtain first audio data representing one or more initial utterances during an interactive voice session with an interactive voice system that interacts with a server to obtain content to provide to a user during the interactive voice session, wherein the interactive voice system is configured to execute on a second computing system and the server is configured to execute on a third computing system;

transmit the first audio data to the interactive voice system via a first communication channel;

generate, based on the first audio data, a prediction that a second audio data representing a subsequent utterance of the user in the interactive voice session will contain sensitive information, wherein the subsequent utterance follows the one or more initial utterances in time;

upon obtaining the second audio data that includes the sensitive information:

transmit third audio data to the interactive voice system via the first communication channel in place of the second audio data, wherein the third audio data represents data that replaces the sensitive information; and

transmit the second audio data to the server via a second communication channel that bypasses the interactive voice system.

15. The one or more non-transitory computer-readable storage media of claim 14 , wherein the instructions further cause the one or more processors of the computing system to:

obtain a timed volatile random subset (t-VRS) for the interactive voice session from the server; and

generate the third audio data based on words in the t-VRS for the interactive voice session.

16. The one or more non-transitory computer-readable storage media of claim 14 , wherein the third audio data is based on a voice of the user.