Systems and methods for managing privacy policies using machine learning
Systems, methods, and devices for managing privacy policies are disclosed. In one embodiment, a method for management of a user's privacy preferences may include: identifying a computer application installed on a user electronic device, or a website accessed using a browser executed by the user electronic device; retrieving a privacy policy document analysis for a privacy policy document associated with the computer application or the website, the privacy policy document analysis comprising a valuation of a plurality of privacy policy segments within the privacy policy document; receiving a privacy preference analysis for the user, the privacy preference analysis comprising a valuation of a plurality of privacy preferences for the user; identifying a recommended action in response to the valuation of one of the privacy policy segments being outside the scope of the valuation of one of the plurality of privacy preferences; and executing the recommended action.
1. A method for management of a user's privacy preferences, comprising:
in an information processing apparatus comprising at least one computer processor:
identifying a computer application installed on a user electronic device, or a website accessed using a browser executed by the user electronic device by examining network traffic to and from the user electronic device via a network monitor separate from the user electronic device;
analyzing a privacy policy document associated with the computer application or the website by:
retrieving historical privacy policy documents for an entity associated with the computer application or website;
segmenting the historical privacy policy documents into a plurality of historical privacy policy segments;
training a neural network to assign sub-category labels to the plurality of historical privacy policy segments, each sub-category label identifying a manner of data collection or identifying data use;
segmenting the privacy policy document into a plurality of privacy policy segments;
predicting a classification for each of the privacy policy segments using the trained neural network; and
generating a valuation of the plurality of privacy policy segments based on the classification;
receiving a privacy preference analysis for the user, the privacy preference analysis comprising a valuation of a plurality of privacy preferences for the user;
identifying a recommended action in response to the valuation of one of the privacy policy segments being outside of a scope of the valuation of one of the plurality of privacy preferences; and
executing the recommended action.
2. The method of claim 1 , wherein each privacy policy segment comprises a privacy category, a privacy sub-category, and a privacy sub-category classification.
3. The method of claim 2 , wherein the valuation of the privacy policy segment is based on at least one of the privacy sub-category and the privacy sub-category classification for the privacy policy segment.
4. The method of claim 1 , wherein the valuation of the privacy policy segment comprises a numeric score, a letter grade, or a color code.
5. The method of claim 1 , further comprising:
modifying the valuation of the privacy policy segment based on an actual implementation of the privacy policy segment.
6. The method of claim 1 , wherein the valuation of a plurality of privacy preference is based on monitored user behavior.
7. The method of claim 1 , wherein the recommended action is to change the valuation of the user privacy preference.
8. The method of claim 1 , wherein the recommended action is based on an action taken by a second user.
9. The method of claim 1 , wherein the recommended action is based on news related to the entity associated with the computer application or website.
10. The method of claim 1 , wherein the recommended action is to restrict use of the computer application or website.
11. The method of claim 1 , wherein the recommended action comprises a restriction on personal identifiable data communicated from the electronic device.
12. The method of claim 1 , wherein the recommended action is generated using a machine learning process.
13. A system for management of privacy preferences for a user, comprising:
an electronic device comprising at least one computer processor;
a privacy policy manager program executed by a server;
a privacy policy manager computer application executed by the electronic device; and
a computer application or browser executed by the electronic device that accessed a website;
wherein:
the privacy policy manager computer application identifies the computer application or the website accessed using the browser by examining network traffic to and from the electronic device via a network monitor separate from the electronic device;
the privacy policy manager computer application retrieves, from the privacy policy manager program, a privacy policy document analysis for a privacy policy document associated with the computer application or the website, the privacy policy document analysis comprising a valuation of a plurality of privacy policy segments within the privacy policy document;
the privacy policy manager computer program analyzes a privacy policy document associated with the computer application or the website by:
retrieving historical privacy policy documents for an entity associated with the computer application or website;
segmenting the historical privacy policy documents into a plurality of historical privacy policy segments;
training a neural network to assign sub-category labels to the plurality of historical privacy policy segments, each sub-category label identifying a manner of data collection or identifying data use;
segmenting the privacy policy document into a plurality of privacy policy segments;
predicting a classification for each of the privacy policy segments using the trained neural network; and
generating a valuation of the plurality of privacy policy segments based on the classification;
the privacy policy manager computer application receives a privacy preference analysis for the user, the privacy preference analysis comprising a valuation of a plurality of privacy preferences for the user;
the privacy policy manager computer application identifies a recommended action in response to the valuation of one of the privacy policy segments being outside of a scope of the valuation of one of the plurality of privacy preferences; and
the privacy policy manager computer application executes the recommended action.
14. The system of claim 13 , wherein the privacy policy manager program generates the privacy policy document analysis using natural language processing of the privacy policy document.
15. The system of claim 13 , wherein each privacy policy segment comprises a privacy category, a privacy sub-category, and a privacy sub-category classification, and the valuation of the privacy policy segment is based on at least one of the privacy sub-category and the privacy sub-category classification for the privacy policy segment.
16. The system of claim 13 , wherein the recommended action is based on at least one of an action taken by a second user and news related to the entity associated with the application or website.
17. The system of claim 13 , wherein the recommended action is to restrict personal identifiable data communicated from the electronic device.
18. The method of claim 1 , wherein the information processing apparatus further analyze a privacy policy document associated with the computer application or the website by training a second machine learning engine for each sub-category label to predict a value for the sub-category label.
19. The system of claim 13 , wherein the privacy policy manager computer application further analyze a privacy policy document associated with the computer application or the website by training a second machine learning engine for each sub-category label to predict a value for the sub-category label.