Work-from-home agent security compliance
Working from home is becoming more and more commonplace. Ensuring that remote workers are following prescribed data protection measures becomes more important, and challenging. Safe and Secure Remote Working Environment (SSRWE) monitors an environment and determines if a non-compliant element is present. For example, a condition of working from home may be that no notetaking or electronic devices, especially ones comprising a camera, be visible to a system camera capturing the field of view in front of a display. If a non-compliant situation is discovered, the display may be redacted and/or other action taken to protect sensitive information.
1. A system, comprising:
a network interface to a network;
a processor configured with machine-readable instructions maintained in a non-transitory storage;
a camera; and
a display presenting information to a user of the system, wherein the information comprises sensitive information;
wherein the processor comprises a neural network trained to recognize an image of a non-compliant environment by presenting a visual notification or an audible notification, wherein the processor receives a camera image from the camera and performs determining that the camera image comprises an image of a non-compliant environment, and wherein operation of the system cannot disable the determination that the camera image comprises the image of the non-compliant environment; and
wherein the processor, upon determining that the camera image does comprise the image of the non-compliant environment, executes a non-compliance response, comprising omitting presentation of the sensitive information by the display.
2. The system of claim 1 , wherein determining whether the camera image comprises the image of the non-compliant environment further comprises a neural network trained to recognize non-compliant environments from the camera image.
3. The system of claim 2 , wherein training the neural network comprises:
accessing a set of object images that indicate the non-compliant environment;
applying one or more transformations to the set of object images, including rotating, color-altering, position translation, contrast alteration, or brightness alteration to create a modified set of object images;
creating a first training set comprising a collected set of object images, the modified set of object images, and a set of compliant environment images;
training the neural network in a first stage using the first training set;
creating a second training set for a second stage of training comprising the first training set and the set of compliant environment images that are incorrectly detected as object images that indicate the non-compliant environment in the first stage of training; and
training the neural network in the second stage of training using the second training set.
4. The system of claim 1 , wherein omitting presentation of the sensitive information by the display comprises disabling the display.
5. The system of claim 1 , wherein omitting presentation of the sensitive information by the display further comprises presenting at least one of a visual notification of omitting the presentation on the display or causing an audio output device to present an audible notification of omitting the presentation.
6. The system of claim 1 , further comprising:
a speaker; and
wherein the processor is further configured to engage in a communication with a customer communication device and receive from a customer, via the customer communication device, sensitive information encoded as audio signals; and
wherein the processor decodes the audio signals and causes the speaker to output the sensitive information therein.
7. The system of claim 6 , further comprising:
an input device to receive machine-inputs from the user; and
wherein the display comprises the sensitive information, transcribed from the speaker's output, and input by the user via the input device.
8. The system of claim 1 , wherein the processor is further configured to obtain from a remote source, via the network, the sensitive information for presentation by the display.
9. The system of claim 1 , wherein the image of the non-compliant environment comprises an image of at least one of a device comprising a camera, a device comprising an audio recorder, a communication device, a writing instrument, a notepad, a non-authorized person, the user engaged in note taking, or an unknown device.
10. The system of claim 1 , wherein the non-compliance response further comprises disabling the system.
11. The system of claim 1 , wherein the non-compliance response comprises alerting, via the network, a communication device of a supervisor.
12. The system of claim 11 , wherein alerting a communication device of the supervisor further comprises automatically connecting the communication device of the supervisor, and wherein at least the camera image is provided to the communication device of the supervisor.
13. The system of claim 1 , further comprising:
at least one of a memory or a storage device; and
wherein the non-compliance response comprises wiping data maintained in the at least one of the memory or the storage device.
14. A method, comprising:
presenting information on a display of a system, wherein the information comprises sensitive information;
receiving a camera image from a camera;
determining, by a process of a processor, that the camera image comprises a neural network trained to recognize an image of a non-compliant environment by presenting a visual notification or an audible notification and wherein operation of the processor cannot disable the determining that the camera image comprises the image of the non-compliant environment; and
upon determining that the camera image does comprise the image of the non-compliant environment, executing a non-compliance response, comprising omitting presentation of the sensitive information by the display.
15. The method of claim 14 , wherein determining whether the camera image comprises the image of the non-compliant environment further comprises a neural network trained to recognize non-compliant environments from the camera image.
16. The method of claim 15 , wherein training the neural network comprises:
accessing a set of object images that indicate the non-compliant environment;
applying one or more transformations to the set of object images, including rotating, color-altering, position translation, contrast alteration, or brightness alteration to create a modified set of object images;
creating a first training set comprising a collected set of object images, the modified set of object images, and a set of compliant environment images;
training the neural network in a first stage using the first training set;
creating a second training set for a second stage of training, comprising the first training set and the set of compliant environment images that are incorrectly detected as object images that indicate the non-compliant environment in the first stage of training; and
training the neural network in the second stage of training using the second training set.
17. The method of claim 14 , wherein omitting presentation of the sensitive information by the display comprises disabling the display.
18. The method of claim 14 , wherein omitting presentation of the sensitive information by the display further comprises presenting at least one of a visual notification of omitting the presentation on the display or causing an audio output device to present an audible notification of omitting the presentation.
19. The method of claim 14 , further comprising:
engaging in a communication with a customer communication device and receiving, from a customer via the customer communication device, sensitive information encoded as audio signals;
decoding the audio signals; and
causing a speaker to output the sensitive information decoded from the audio signals.
20. A system, comprising:
a network interface to a network;
a processor configured with machine-readable instructions maintained in a non-transitory storage;
a microphone; and
a display presenting information to a user of the system, wherein the information comprises sensitive information;
wherein the processor comprises a neural network trained to recognize an image of a non-compliant environment by presenting a visual notification or an audible notification, wherein the processor receives a sound from the microphone and determines that the sound comprises a sound portion from a non-compliant environment and wherein operation of the system cannot disable the determination that the sound comprises the sound portion from the non-compliant environment; and
upon determining that the sound portion does comprise sound from the noncompliant environment, the processor performs executing a non-compliance response, comprising omitting presentation of the sensitive information by the display.