Selective artificial intelligence processing of sensitive data
Methods, systems, and apparatuses are described herein for efficiently identifying and processing sensitive data. A computing device may receive text content, such as one or more words. The computing device may select characters of that text content by identifying, using regular expressions corresponding to sensitive data categories, matches and select various characters including and around such matches. The computing device may provide the selected characters as input to a machine learning model trained to identify sensitive data, and that trained machine learning model may output information about a sensitivity of the characters. Such output might be used to modify all or portions of the text content.
1 . A first computing device comprising:
one or more processors; and
memory storing instructions that, when executed by the one or more processors, cause the first computing device to:
receive, from a second computing device, text content;
test each of a plurality of different portions of the text content for sensitive data;
based on a determination that a first portion of the plurality of different portions of the text content comprises first sensitive data:
determine a sensitive data category of the first sensitive data;
provide, based on the sensitive data category of the first sensitive data and as input to a first trained machine learning model, a first plurality of characters of the first portion of the plurality of different portions of the text content; and
generate, based on output from the first trained machine learning model, a first token;
based on a determination that a second portion of the plurality of different portions of the text content comprises second sensitive data:
determine a sensitive data category of the second sensitive data; and
generate, based on the sensitive data category of the second sensitive data, a second token by processing a second plurality of characters of the second portion of the plurality of different portions of the text content using a tokenization algorithm;
generate a tokenized version of the text content by:
replacing the first portion of the plurality of different portions of the text content with the first token; and
replacing the second portion of the plurality of different portions of the text content with the second token;
receive, from a third computing device, a request for the text content; and
send, to the third computing device and in response to the request, the tokenized version of the text content.
2 . The first computing device of claim 1 , wherein the instructions, when executed, further cause the first computing device to:
identify the first trained machine learning model from a plurality of trained machine learning models based on the sensitive data category of the first sensitive data.
3 . The first computing device of claim 1 , wherein the instructions, when executed, further cause the first computing device to generate the second token by encrypting at least a portion of the second plurality of characters.
4 . The first computing device of claim 1 , wherein the instructions, when executed, further cause the first computing device to test each of the plurality of different portions of the text content for sensitive data by comparing at least a portion of the text content to a regular expression.
5 . The first computing device of claim 1 , wherein the determination that a first portion of the plurality of different portions of the text content comprises first sensitive data is based on an identification, in the first portion, of one or more words associated with the sensitive data category of the first sensitive data.
6 . The first computing device of claim 1 , wherein the instructions, when executed, further cause the first computing device to:
receive, from the third computing device, a request for an original form of the text content; and
send, based on authenticating the third computing device, the text content to the third computing device.
7 . The first computing device of claim 1 , wherein the instructions, when executed, further cause the first computing device to generate the second token based on a format of the second plurality of characters.
8 . A method comprising:
receiving, by a first computing device and from a second computing device, text content;
testing each of a plurality of different portions of the text content for sensitive data;
based on a determination that a first portion of the plurality of different portions of the text content comprises first sensitive data:
determining a sensitive data category of the first sensitive data;
providing, based on the sensitive data category of the first sensitive data and as input to a first trained machine learning model, a first plurality of characters of the first portion of the plurality of different portions of the text content; and
generating, based on output from the first trained machine learning model, a first token;
based on a determination that a second portion of the plurality of different portions of the text content comprises second sensitive data:
determining a sensitive data category of the second sensitive data; and
generating, based on the sensitive data category of the second sensitive data, a second token by processing a second plurality of characters of the second portion of the plurality of different portions of the text content using a tokenization algorithm;
generating a tokenized version of the text content by:
replacing the first portion of the plurality of different portions of the text content with the first token; and
replacing the second portion of the plurality of different portions of the text content with the second token;
receiving, from a third computing device, a request for the text content; and
sending, to the third computing device and in response to the request, the tokenized version of the text content.
9 . The method of claim 8 , further comprising:
identify the first trained machine learning model from a plurality of trained machine learning models based on the sensitive data category of the first sensitive data.
10 . The method of claim 8 , wherein the generating the second token comprises encrypting at least a portion of the second plurality of characters.
11 . The method of claim 8 , wherein the testing each of the plurality of different portions of the text content for sensitive data comprises comparing at least a portion of the text content to a regular expression.
12 . The method of claim 8 , wherein the determination that a first portion of the plurality of different portions of the text content comprises first sensitive data is based on an identification, in the first portion, of one or more words associated with the sensitive data category of the first sensitive data.
13 . The method of claim 8 , further comprising:
receiving, from the third computing device, a request for an original form of the text content; and
sending, based on authenticating the third computing device, the text content to the third computing device.
14 . The method of claim 8 , wherein the generating the second token is based on a format of the second plurality of characters.
15 . One or more non-transitory computer-readable media storing instructions that, when executed by one or more processors of a first computing device, cause the first computing device to:
receive, from a second computing device, text content;
test each of a plurality of different portions of the text content for sensitive data;
based on a determination that a first portion of the plurality of different portions of the text content comprises first sensitive data:
determine a sensitive data category of the first sensitive data;
provide, based on the sensitive data category of the first sensitive data and as input to a first trained machine learning model, a first plurality of characters of the first portion of the plurality of different portions of the text content; and
generate, based on output from the first trained machine learning model, a first token;
based on a determination that a second portion of the plurality of different portions of the text content comprises second sensitive data:
determine a sensitive data category of the second sensitive data; and
generate, based on the sensitive data category of the second sensitive data, a second token by processing a second plurality of characters of the second portion of the plurality of different portions of the text content using a tokenization algorithm;
generate a tokenized version of the text content by:
replacing the first portion of the plurality of different portions of the text content with the first token; and
replacing the second portion of the plurality of different portions of the text content with the second token;
receive, from a third computing device, a request for the text content; and
send, to the third computing device and in response to the request, the tokenized version of the text content.
16 . The one or more non-transitory computer-readable media of claim 15 , wherein the instructions, when executed, further cause the first computing device to:
identify the first trained machine learning model from a plurality of trained machine learning models based on the sensitive data category of the first sensitive data.
17 . The one or more non-transitory computer-readable media of claim 15 , wherein the instructions, when executed, further cause the first computing device to generate the second token by encrypting at least a portion of the second plurality of characters.
18 . The one or more non-transitory computer-readable media of claim 15 , wherein the instructions, when executed, further cause the first computing device to test each of the plurality of different portions of the text content for sensitive data by comparing at least a portion of the text content to a regular expression.
19 . The one or more non-transitory computer-readable media of claim 15 , wherein the determination that a first portion of the plurality of different portions of the text content comprises first sensitive data is based on an identification, in the first portion, of one or more words associated with the sensitive data category of the first sensitive data.
20 . The one or more non-transitory computer-readable media of claim 15 , wherein the instructions, when executed, further cause the first computing device to:
receive, from the third computing device, a request for an original form of the text content; and
send, based on authenticating the third computing device, the text content to the third computing device.