Warning: Constant ABSPATH already defined, this will be an error in PHP 9 in /home/unhinder/www/www/Phi/wp-config.php on line 22
Computing – Page 87 – I am … unhindered by talent

The scary power of search logs


Warning: Undefined array key "HTTP_REFERER" in /home/unhinder/www/www/Phi/wp-content/plugins/google-hilite.php on line 58

Stay off those rails
My experience is that it’s often hard to get people excited about things like Google’s highly detailed records of what you’ve been searching for (e.g., check out “Verbatim: Search firms surveyed on privacy” for a survey of what four major search engines collect, “FAQ: When Google is not your friend” for further analysis, and “Keeping secrets” on Slate for a related discussion). Responses often run along the “I have nothing to hide” or “Only criminals/bad people would have reason to worry about that” vein, but an underlying theme is often a sense that there’s really nothing useful/interesting to be learned from trolling through search terms.

Recently, however, there was a leak of the search histories of over half a million AOL users over a three month period. No names are attached, but they do have user numbers so you can pull together all the search terms from single individuals. Declan McCullagh (of the Politech mailing list) has sifted through some of it in “AOL’s disturbing glimpse into users’ lives” over on CNet, and it’s a real eye opener. And while there are clearly some deeply scary and deeply troubled people in his discussion (makes for better reading), step back for a moment from the pathologies and think about just how much these small subsets of these people’s search histories tell us about them and their lives. And then think about how much of your life (good, bad, and slightly smelly) could be reconstructed from your search history. Then throw in some web access logs, your Amazon search and buying patterns, your e-mail address book, and IM logs.

To misquote a fairly paranoid TV show: The data are out there.

And they can be accidentally leaked, deliberately sold (with or without the knowledge of the holding company), or subpoenaed. Currently there is very little legal protection of your privacy in the realm of search terms; basically all you’ve got is the good will and good word of the search engines, and faith that the courts care and get it enough to look after your privacy interests. As the AOL situation makes clear, though, mistakes happen, and lord knows that not all judges are equally together on these issues.

What to do? Not clear that there’s a simple “fix it” action here, but at a minimum it would be useful to spread the word and help raise awareness. Bug your favorite search engine and let them know that you have expectations of privacy and are willing to vote with your feet. Supporting the work of groups like the Center for Democracy and Technology and the ACLU wouldn’t hurt, since they’ve both been pretty front-line on these and other important privacy issues. You might also take a librarian to lunch, since they’re a pretty cool bunch when it comes to protecting the privacy of their users.

The more hard-core among you might explore the many anonymous browsing tools. These are usually just a server that sits between you and, for example, Google, so Google’s records show all the searches for lots of people (including you) as coming from the anonymous proxy. This sounds good in principle, but in some sense you’re just moving your trust from Google to the proxy, since their logs can allow people to reconstruct your search (and browsing) history. If you’ve got toys, you could set up your own proxy (you do trust yourself, don’t you?), but then you lose the anonymity that comes from mixing your history in with that of hundreds of other uses. In other words, this is a complicated road to travel, and you’d best do your homework if you really want to anonymize your history.

Be careful out there. And don’t talk to strangers (at least not too often)…

Our ability to internalize a manipulated world


Warning: Undefined array key "HTTP_REFERER" in /home/unhinder/www/www/Phi/wp-content/plugins/google-hilite.php on line 58

Psychedelic music
Over on Notional Slurry, Bill Tozier (who’s one of the many cool people I get to talk to here at GECCO) uses a funky anomoly in a Google map image to raise a very interesting question:

Google Book Search will become a stunning addition to the armamentarium, but is full of first drafts of scanned books: full of errors and missing pages.

Will we all go through Google Book Search and replace those missing and damaged digitized pages, or will we adapt to their lack by simply calling it good enough?

Way back in the pre-dawn of history (i.e., when I was an undergrad and we still copied LPs onto cassette tape), I made a copy of a friend’s (“Hi Jean!”) LP of Vivaldi’s Four Seasons. Unfortunately my cassette was a hair (2 minutes at most) shorter than the music, so my copy was missing those last few bars.

About three years later I was studying math in Budapest, and a bunch of us went to a (wonderful) performance of Four seasons. After many minutes of thoroughly enjoying the experience, all of a sudden something just felt wrong, and shortly after that the piece ended. As everyone stood and applauded, I figured out what the “problem” was: In my head, the piece was supposed to have ended a minute or two earlier! I’d listened to that tape enough that I’d internalized a finish that I intellectually knew was not the actually ending. The live performance then became “wrong” or “broken” when the orchestra didn’t abruptly cease playing at the point where (in my head) the tape ran out. (Imagine the chaos and confusion in the hall if they had!)

As a much more significant example, consider the errors and modifications (often unintentional, but no less serious as a result) introduced in the millenia of hand copied texts, including such politically and culturally powerful documents as the bible. Here we have mistakes and changes becoming internalized to the point of becoming church policy, cultural norms, and sometimes even civil law!

So obviously we can adapt all too well to errors and manipulation, and these things can become the new “truth” all too easily. It would thus seem crucial that we be vigilant in looking for these errors (especially in potentially powerful repositories like Google Books) and expect (even demand!) that these things be fixed when they’re brought to the attention of Google (or Amazon or libraries or …). This is one place where a public resource like Wikipedia has an advantage over a private resource like Google Books: If I see an error in a Wikipedia entry, I can fix it myself without having to rely on someone else to actually make the change for me.

After all this carping, though, it’s worth returning to the title of Bill’s post: “Bill in Escherland”. Much great art is great in significant part because it breaks and reassembles your conception of art, information, and the world. While I would hardly call my copy of Four Seasons art, the effect it had on me is quite similar in many ways to how sampling and remixes can help/force us to perceive ideas in a new way. Viva la Escher!