Preventing of audio attacks

• US 10,134,396 B2
• Filed: 12/07/2016
• Issued: 11/20/2018
• Est. Priority Date: 12/07/2016
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google LLC (Alphabet Inc.)
• Inventors: Campbell, Lee, Beder, Samuel Kramer
• Primary Examiner(s): Kazeminezhad, Farzad
• Application Number:
  US15/371,929
• Publication Number:
  US 20180158454A1
• Time in Patent Office: 713 Days
• Field of Search: 713186, 705 39
• CPC Class Codes: G06Q 20/32 using wireless devices
• CPC Class Codes: G06Q 20/40145 Biometric identity checks
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications
• CPC Class Codes: G10L 2015/223 Execution procedure of a spoken command
• CPC Class Codes: H04L 51/18 Commands or executable codes
• CPC Class Codes: H04L 51/222 using geographical location information, e.g. messages transmitted or received in proximity of a certain spot or area

Abstract

In some implementations, a method is performed by computing devices, and includes receiving a first message indicating that an improper voice command has been detected by a first device, identifying a user account associated with the first device based on the first message, identifying a second device associated with the user account, and, in response to receiving the first message indicating that the voice command detected by the first device is improper, sending a second message to the second device, the second message indicating that the voice command should not be performed.

Claims

1. A method comprising:

receiving, over a network by a server and from a first device, a first message, where the first message includes encrypted audio data that indicates that the first device detected that an improper voice command was audibly output by the first device;

in response to receiving the first message that indicates that the first device detected that the improper voice command was audibly output by the first device;

identifying, by the server, a user account associated with the first device based on the first message;

identifying, by the server, a second device associated with the user account; and

sending, over the network by the server to the second device, a second message, where the second message indicates that a voice command corresponding to the improper voice command that is captured by a microphone of the second device should not be performed.

2. The method of claim 1, wherein sending, over the network by the server to a second device, a second message, where the second message indicates that a corresponding voice command that is captured by a microphone of the second device should not be performed comprises:

receiving, by the server, audio data for the corresponding voice command captured by the second device from the second device;

determining, by the server, that the audio data for the corresponding voice command captured by the second device corresponds to audio data for the improper voice command detected by the first device; and

in response to determining that the audio data for the corresponding voice command captured by the second device corresponds to the audio data for the improper voice command detected by the first device, not executing the corresponding voice command captured by the second device and sending the second message to the second device.

3. The method of claim 2, wherein determining that the audio data for the corresponding voice command captured by the second device corresponds to the audio data for the improper voice command detected by the first device comprises:

determining, by the one or more computing devices, that the first message and the audio data for the corresponding voice command captured by the second device are associated with the user account associated with the first device.

4. The method of claim 2, wherein determining that the audio data for the corresponding voice command captured by the second device corresponds to the audio data for the improper voice command detected by the first device comprises:

determining, by the server, that a first time at which the audio data for the improper voice command detected by the first device was detected and a second time at which the audio data for the corresponding voice command was captured by the second device are within a certain period of time.

5. The method of claim 2, wherein determining that the audio data for the corresponding voice command captured by the second device corresponds to the audio data for the improper voice command detected by the first device comprises:

determining, by the server, that the audio data for the corresponding voice command captured by the second device and the audio data for the improper voice command detected by the first device have a minimum level of similarity.

6. The method of claim 2, wherein determining that the audio data for the corresponding voice command captured by the second device corresponds to the audio data for the improper voice command detected by the first device comprises:

determining, by the server, that the audio data for the corresponding voice command captured by the second device and the audio data for the improper voice command detected by the first device request a same type of action.

7. The method of claim 2, wherein determining that the audio data for the corresponding voice command captured by the second device corresponds to the audio data for the improper voice command detected by the first device comprises:

performing, by the server, automatic speech recognition on the audio data for the corresponding voice command captured by the second device to generate a first processed audio segment;

performing, by the server, automatic speech recognition on the audio data for the improper voice command detected by the first device to generate a second processed audio segment; and

determining, by the server, that the first processed audio segment and the second processed audio segment have a minimum level of similarity.

8. The method of claim 1, further comprising:

determining, by the server, a set of multiple devices associated with the user account based on at least one of;

a location of each device of the set of the multiple devices, a label generated by a user of the user account, or a grouping generated by the user; and

sending, by the server, a third message to each device of the set of multiple devices.

9. The method of claim 8, wherein the set of multiple devices associated with the user account is a proper subset of the multiple devices associated with the user account, and wherein each device of the set of multiple devices is selected for inclusion in the set of multiple devices based on at least one of:

a location of each device of the set of the multiple devices, a label generated by a user of the user account, or a grouping generated by the user.

10. The method of claim 8, further comprising:

determining, by the server, a location of the first device based on at least one of;

device registration data previously stored by the server, network connectivity data, or location data sent by the first device;

identifying, by the server and based on the location of the first device, a set of multiple devices associated with the user account, each device of the set of multiple devices having a location within a maximum distance of the location of the first device; and

sending, by the server, a third message to each device of the set of multiple devices, the third message indicating that the corresponding voice command should not be performed.

11. The method of claim 1, wherein the first message indicates that the improper voice command was falsified.

12. The method of claim 1, wherein the second message to the second device prevents execution of the corresponding voice command.

13. The method of claim 1, wherein identifying, by the server, a second device associated with the user account comprises:

determining, by the server, a location of the first device based on at least one of;

device registration data previously stored by the server, network connectivity data, or location data sent by the first device; and

identifying, by the server and based on the location of the first device, the second device.

14. The method of claim 1, further comprising:

in response to receiving the first message, preventing, by the server, voice commands from being executed by the server for a period of time after receiving the first message.

15. The method of claim 1, wherein receipt of the second message by the second device triggers the second device to, for a period of time, block processing of voice commands in audio captured by the microphone of the second device.

16. The method of claim 1, wherein the second message indicates a time corresponding to detection of the improper voice command and the second device is triggered by the second message to ignore any voice command detected in audio captured by the microphone of the second device within a predetermined time period of the time corresponding to detection of the improper voice command.

17. The method of claim 1, wherein the first message comprises encrypted audio data, where the encrypted audio data includes an indication that the first device audibly output an improper voice command.

18. A system comprising:

one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations comprising;

receiving, over a network by a server and from a first device, a first message, where the first message includes encrypted audio data that indicates that the first device detected that an improper voice command was audibly output by the first device;

in response to receiving the first message that indicates that the first device detected that the improper voice command was audibly output by the first device;

identifying, by the server, a user account associated with the first device based on the first message;

identifying, by the server, a second device associated with the user account; and

sending, over the network by the server to the second device, a second message, where the second message indicates that a voice command corresponding to the improper voice command that is captured by a microphone of the second device should not be performed.

19. A computer-readable storage device storing software comprising instructions executable by one or more computers which, upon such execution, cause the one or more computers to perform operations comprising:

receiving, over a network by a server and from a first device, a first message, where the first message includes encrypted audio data that indicates that the first device detected that an improper voice command was audibly output by the first device;

in response to receiving the first message that indicates that the first device detected that the improper voice command was audibly output by the first device;

identifying, by the server, a user account associated with the first device based on the first message;

identifying, by the server, a second device associated with the user account; and

sending, over the network by the server to the second device, a second message, where the second message indicates that a voice command corresponding to the improper voice command that is captured by a microphone of the second device should not be performed.