Centralized and decentralized data protection through redaction

1 . A system comprising:

at least one machine learning model;

one or more processors; and

computer memory storing computer-usable instructions that, when executed by the one or more processors, perform operations comprising:

detecting, by using the at least one machine learning model, an entry of sensitive data at a sensitive field on an active user interface;

based on detecting the entry of the sensitive data at the sensitive field, redacting at least a portion of the sensitive data as the sensitive data is being received on the active user interface;

tokenizing the sensitive data such that the sensitive data is tokenized prior to being transmitted;

receiving, at an application programming interface (API) gateway or the router, a request from a user interface to unmask at least a portion of the sensitive data;

validating permissions for unmasking the at least portion of the sensitive data based on accessing an active directory;

receiving approval and a decryption token from an enterprise token repository in response to the approval; and

permitting un-redaction of at least a portion of the sensitive data via an un-redact button on a navigation menu, wherein a record of the un-redaction is stored at an offline database.

2 . The system of claim 1 , further comprising detecting the entry of the sensitive data at the sensitive field using natural language processing.

3 . The system of claim 1 , further comprising determining that a website corresponding to the active user interface and the sensitive field is malicious by comparing a URL and letterhead associated with the website.

4 . The system of claim 3 , further comprising preventing, by a network router and based on determining that the website is malicious, transmission of the sensitive data prior to transmitting the sensitive data from an application programming interface gateway.

5 . The system of claim 4 , further comprising transmitting an alert to a user device corresponding to the entry of the sensitive data at the sensitive field, wherein the alert indicates that the website is malicious.

6 . The system of claim 1 , further comprising:

causing the tokenized sensitive data to be transmitted to a first endpoint within a network that includes a network layer; and

causing the tokenized sensitive data to be transmitted from the first endpoint to a second endpoint within the network.

7 . The system of claim 1 , further comprising validating, based on user permissions, a request to un-redact at least a portion of the sensitive data; and

based on the validation, un-redacting the at least portion of the sensitive data for display on the active user interface.

8 . The system of claim 1 , wherein the sensitive data is detected based on listening to keystrokes on the active user interface and identifying, by the at least one machine learning model, a pattern of the keystrokes.

9 . The system of claim 1 , further comprising determining that a website corresponding to the active user interface is malicious based on an IP destination address not having access to an edge node of a network, and preventing transmission of the sensitive data from a network router prior to transmitting the sensitive data from the application programming interface gateway.

10 . The system of claim 1 , wherein detecting the entry of the sensitive data comprises using natural language processing to detect a medical professional identifier having a prefix or suffix associated with an individual's name, wherein the prefix or suffix comprises at least one of “Dr.”, “MD”, or “RN”.

11 . A computerized method for centralized and decentralized data protection, the method comprising:

detecting, by an application programming interface (API) gateway using at least a machine learning model trained using a plurality of application programming interface requests, that data being received by a router includes sensitive data;

based on detecting that the data being received by the router includes the sensitive data, applying, by the API gateway, a masking technique comprising tokenization to at least a portion of the sensitive data prior to transmitting the sensitive data from the API gateway to a network layer;

transmitting, by the API gateway, at least the portion of the masked sensitive data to the network layer;

receiving, at an application programming interface (API) gateway or the router, a request from a user interface to unmask at least a portion of the sensitive data;

validating permissions for unmasking the at least portion of the sensitive data based on accessing an active directory;

receiving approval and a decryption token from an enterprise token repository in response to the approval; and

permitting un-redaction of at least a portion of the sensitive data via an un-redact button on a navigation menu, wherein a record of the un-redaction is stored at an offline database.

12 . The computerized method of claim 11 , wherein the masking technique comprises redacting the sensitive data from display on an active user interface of a user device.

13 . The computerized method of claim 11 , further comprising:

receiving the sensitive data from an application downloaded on at least one user device, wherein the application is managed by a container orchestration platform; and

detecting that the data being received by the router includes the sensitive data based on an application programming interface request, from the application, for a security token for encryption.

14 . The computerized method of claim 11 , wherein the sensitive data is detected based on listening to keystrokes on at least one user device and identifying, via the machine learning model, a pattern of the keystrokes.

15 . The computerized method of claim 11 , further comprising: detecting, by using the machine learning model, that sensitive data is being entered at a sensitive field within a user interface; encrypting one or more attributes of the sensitive data based on a type of sensitive data being encrypted; and requesting a security token from an enterprise token repository for encrypting the sensitive data prior to transmitting the sensitive data across a layer of a network.

16 . The computerized method of claim 11 , wherein the machine learning model is deployed on a Kubernetes master node that employs sensitive data detection rules to orchestrate worker nodes and pods of a Kubernetes cluster, wherein the worker nodes and pods are associated with a network provider, and wherein the Kubernetes master node detects the sensitive data based on an input received at a sensitive field on an active user interface.

17 . Non-transitory computer-readable media having computer-usable instructions embodied thereon that, when executed by a processor, perform operations for centralized and decentralized data protection, the operations comprising:

receiving, by a router and from a user device, sensitive data;

detecting, by using at least a machine learning model, the sensitive data based on an entry of the sensitive data at the user device and based on receiving the sensitive data at the router;

based on detecting the sensitive data, applying a masking technique comprising tokenization to at least a portion of the sensitive data received prior to transmitting the sensitive data to a network layer within a network, such that the sensitive data is masked once it enters the network layer and as it is transmitted from the network layer;

receiving, at an application programming interface (API) gateway or the router, a request from a user interface to unmask at least a portion of the sensitive data;

validating permissions for unmasking the at least portion of the sensitive data based on accessing an active directory;

receiving approval and a decryption token from an enterprise token repository in response to the approval; and

permitting un-redaction of at least a portion of the sensitive data via an un-redact button on a navigation menu, wherein a record of the un-redaction is stored at an offline database.

18 . The non-transitory computer-readable media of claim 17 , wherein the sensitive data is detected based on the entry of the sensitive data within a sensitive field of an active user interface.

19 . The non-transitory computer-readable media of claim 17 , wherein the sensitive data is detected based on an application programming interface request, from an application corresponding to the user device, for a security token for encryption.