Real-time detection of privacy leaks based on linguistic features

1. A method for detecting privacy leaks, the method being performed by at least one computing device having at least one processor, the method comprising:

receiving, at the computing device from a client device, a keyword;

generating, by the computing device, a first query comprising the keyword;

determining, by the computing device, at least one word representing a sensitive topic associated with the keyword;

generating, by the computing device, a second query comprising the keyword logically combined with the at least one word;

transmitting, from the computing device to a search engine, the first query and the second query;

receiving, at the computing device from the search engine, a number of search results found by the search engine for the first query and the second query;

determining, by the computing device, a confidence score for the keyword based on the number of search results for the first query and the second query, the confidence score indicating a probability that the keyword would leak information associated with the sensitive topic; and

transmitting the confidence score from the computing device to the client device.

2. The method of claim 1 , further comprising, when the confidence score meets or exceeds a predetermined threshold, performing a security action to prevent a leak of the information associated with the sensitive topic.

3. The method of claim 2 , wherein the security action comprises transmitting, from the computing device to the client device, one or more of:

an indication to remove the keyword from text entered into the client device;

an indication to prevent the keyword from being entered into subsequent text on the client device; and

an indication to display a warning about the keyword on the client device.

4. The method of claim 1 , further comprising:

determining, by the computing device based on a word hierarchy, a replacement word for the keyword; and

transmitting, from the computing device to the client device, the replacement word.

5. The method of claim 4 , wherein determining the replacement word comprises:

identifying the keyword in the word hierarchy;

locating a parent word of the keyword in the word hierarchy, the parent word being at a different level in the word hierarchy than the keyword; and

selecting the parent word as the replacement word.

6. The method of claim 1 , further comprising:

prior to transmitting the first query and the second query to the search engine, determining, by the computing device, whether a cached results database includes a cached result for the first query.

7. The method of claim 6 , further comprising:

transmitting the cached result for the first query from the computing device to the client device if the cached results database includes the cached result for the first query.

8. The method of claim 1 , wherein receiving the keyword from the client device comprises receiving the keyword from a privacy leakage plug-in executing within an email client of the client device.

9. The method of claim 1 , wherein the second query is generated to return search results from the search engine for the sensitive topic.

10. The method of claim 1 , further comprising:

determining, by the computing device, at least one other word representing another sensitive topic associated with the keyword;

generating, by the computing device, a third query comprising the keyword logically combined with the at least one other word;

transmitting, from the computing device to the search engine, the third query;

receiving, at the computing device from the search engine, a number of additional search results found by the search engine for the third query;

determining, by the computing device, an additional confidence score for the keyword based on the number of additional search results for the third query, the additional confidence score indicating a probability that the keyword would leak information associated with the other sensitive topic;

ranking, by the computing device, the confidence scores to generate a tuple of relevant topics; and

transmitting the tuple of relevant topics from the computing device to the client device.

11. A computing device comprising:

a processor; and

a memory coupled to the processor and having instructions stored thereon which, when executed by the processor, perform operations for detecting privacy leaks, the operations comprising:

receiving, at the computing device from a client device, a keyword;

generating, by the computing device, a first query comprising the keyword;

determining, by the computing device, at least one word representing a sensitive topic associated with the keyword;

generating, by the computing device, a second query comprising the keyword logically combined with the at least one word;

transmitting, from the computing device to a search engine, the first query and the second query;

receiving, at the computing device from the search engine, a number of search results found by the search engine for the first query and the second query;

determining, by the computing device, a confidence score for the keyword based on the number of search results for the first query and the second query, the confidence score indicating a probability that the keyword would leak information associated with the sensitive topic; and

transmitting the confidence score from the computing device to the client device.

12. The computing device of claim 11 , wherein the operations further comprise, if the confidence score meets or exceeds a predetermined threshold, performing a security action to prevent a leak of the information associated with the sensitive topic.

13. The computing device of claim 11 , wherein the operations further comprise:

determining, by the computing device based on a word hierarchy, a replacement word for the keyword; and

transmitting, from the computing device to the client device, the replacement word.

14. A method for displaying privacy leak information, the method being performed by at least one computing device having at least one processor, the method comprising:

displaying, on a display device of the computing device, a user interface (UI) capable of receiving text from an input device associated with the computing device;

receiving, through the UI, text including a plurality of words;

identifying, by the computing device, a keyword among the plurality of words;

transmitting, from the computing device to a server, the keyword;

receiving, at the computing device from the server, a confidence score associated with the keyword and a replacement word for the keyword; and

updating the UI to include the confidence score and the replacement word.

15. The method of claim 14 , wherein:

the confidence score indicates a first probability that the keyword would leak information for a sensitive topic associated with the keyword; and

a second probability that the replacement word would leak information for the sensitive topic associated with the keyword is less than the first probability.

16. The method of claim 15 , wherein:

the method further comprises, when the confidence score meets or exceeds a predetermined threshold, performing a security action to prevent a leak of the information associated with the sensitive topic; and

the security action comprises one or more of:

removing the keyword from the text;

preventing the keyword from being entered into subsequent text; and

displaying a warning about the keyword.

17. The method of claim 14 , further comprising:

receiving, through the UI, a selection of the replacement word; and

replacing, in the UI, the keyword with the replacement word in the text.

18. The method of claim 14 , further comprising displaying, in the UI, an indication that the keyword potentially leaks sensitive information.

19. The method of claim 18 , wherein the indication that the keyword potentially leaks sensitive information includes a color-coding of the keyword in the UI.

20. The method of claim 14 , further comprising:

prior to updating the UI to include the confidence score and the replacement word, receiving a selection of the keyword; and

displaying a privacy report associated with the selected keyword in the UI.