Google crawls and tracks nearly everything from every user to every click and every social link. They probably know yours friends connectivity three, six, perhaps a dozen times deep. That's the fiends of friends of friends of friends of friends of friends of friends of friends of friends of friends of friends of your friend. Perhaps. They know what you buy online, what you search for, click on, linger on, how you got there and maybe more. Facebook collects some personal information and lots of public information. The sell the use of API's for advertisers can track you with the use of resistant cookies, super cookies, fingerprints (MAC & IP addresses), and who knows what else. They also know the same info about your friends, and the links you may not even know about. The Google API's mashed up with Facebook's, Twitter's, flickr's, etc., etc., etc. can match user names, groups, affiliations, with individual activity. It just takes the time and effort along with sufficient skills and motivation.
Google once proclaimed on their public blog that they were proud to have hired to top-level programmers away from the CIA. Really? There was a rumor that the CIA had an office on the same campus as the Google complex, and even shared a wall. Probably not true, but that was the rumor that I had heard in three different web sites at the time. Hard to believe. Even scary to think it could be true.
Is it possible that could could publicly fight tooth & nail against disclosing any personal data without a wiretap warrant and at the same time, perhaps unwittingly (denial is a powerful adversary), be providing detailed personal data to a government agency? Maybe the CIA is just a platinum class member of advertisers, who may have storefront websites set-up as a rouse. They can buy all the API's they want from the search engines and social websites or create their own API's to collect data with social engineering, advertising and super-persistent cookies. They may be able to garner information on perhaps 90% of all web users, with semi-publicly available computer codes and databases.
Just what is the worst case scenario possible today or next year? Assuming someone pulled all the resources available from the top five of each of the most populace social websites (via advertiser's tools) together, with advertisers access to Google data, and performed a remarkably efficient mash-up of scripted API's and created their own coherent database of records for individuals with links to all of their associations. How much information can they cull for each individual? What could they tell me about me? How scary is that?