Comments on: WSJ Hosts Search History/Privacy Debate

By: Kamal Jain

Kamal Jain — Mon, 21 Aug 2006 03:26:49 +0000

Jakob, your proposal for search engines to defend against the noise can't really work. It will be quite dangerous to put nonsense results among the top results. Doing so will give impression of lower quality when humans do the searches. So all this privacy defending utility has to do is to click a link among the top results.

Markets depends upon tug-of-war game to improve the human lives. If enough users do fake searches to defend their privacy, it puts pressure on search engines to voluntary start respecting privacy.

Kamal Jain

Disclaimer: The author is an employee in Microsoft. The opinion expressed is his own personal opinion.

By: Jakob Nielsen

Jakob Nielsen — Sun, 20 Aug 2006 23:39:51 +0000

Noise in the search logs is a great way to enhance privacy, but it would place a big burden on the search engines. To work well, the majority of searches from a user's machine would have to be software-generated, and the privacy utility would have to simulate a human user, including behaviors like clicking on both organic search results and ads.

Clickfraud (from the advertiser's perspective) would explode, when, say, 80% of clicks are really software-generated.

Luckily users don't tend to employ natural language when constructing search queries, so I think it shuld be possible to write software that would generate queries that would be hard to differentiate from human-written ones. Much simpler than passing the Turing test. But still not completely easy.

Here's a possible counter-measure for the search engines: seed a few SERPs with a nonsense result, and if it's "clicked" on, there's a big probability that you are talking to a computerized privacy utility ratner than a human (since users tend to click on search headlines that make sense).

By: Kamal Jain

Kamal Jain — Fri, 18 Aug 2006 22:00:58 +0000

Actually there are couple of standard methods for individuals to protect their privacy. One is to not accept any unnecessary cookie. Two, install a desktop component to add some noise in your search log. Once in a while this desktop component could make a random search on the search engine you regularly use.

I am not a firefox fan, but there might be something like this already available as a firefox extension.

By: Kamal Jain

Kamal Jain — Fri, 18 Aug 2006 21:54:58 +0000

Do not forget that Google owns 5% of AOL. So 5% accountability of AOL's misshap falls on Google's shoulders too.

An accident is an accident. No body could claim, no matter how safe driver he/she is, that an accident will not happen again. In Eric Schmidt's word, never say never.

Even if I assume that the chances of a privacy accident at Google is small, but the impact could be so huge that the tradeoff is perhaps not worth while.

By: JG

JG — Fri, 18 Aug 2006 19:48:38 +0000

To follow up on my own post: In further support of this idea of my query stream as a "performance", the Germans have a word: Lebenskuenstler (literally: "life artist"). The notion is someone who lives their life as a work of art; the artist's medium is not a canvas or a block of marble, but the artist's own life. One's actions and intentions are one's artistic expression and creation.

So if I choose to define myself as a Lebenskuenstler, I truly feel that companies should not have the right to reproduce my actions and intentions, the works of my life art.

There is, I suppose, still a question of ownership of the digital data. Do I own it, because I am the creator? Or does G/Y/M own it, because I've "expressed" it to them? While personally I feel that I should still have 100% ownership and that I am only "licensing" my Lebenskunst to G/Y/M (the same way a radio broadcast does not give me ownership of someone else's music), I do see how one could make the point that the search engines do at least have partial ownership.

But even with partial ownership, they still do not have the right to re-publish my "art", because I still retain the other part of the ownership. I'm sorry G/Y/M: No third parties allowed. The terms and conditions of my Lebenskunst do not permit it.

It is about time that the DMCA start working for the little guy.

By: Frank

Frank — Fri, 18 Aug 2006 16:40:22 +0000

JG's point is very interesting. I recommend taking a look at Justin Hughes's article on ssrn on whether "size matters" in copyright law.

I've got a post on the "microworks" issue here; Grimmelmann's comment is classic:

http://prawfsblawg.blogs.com/prawfsblawg/2006/05/overheard_in_ne.html

By: Chetan

Chetan — Fri, 18 Aug 2006 16:03:47 +0000

I agree. I have said before Privacy is Google’s Achilles Heel just like Security has been for Microsoft. MS ignored Security for a long time until they couldn’t.

By: King Troll

King Troll — Fri, 18 Aug 2006 13:13:11 +0000

I needed money for smokes, one of my friends knew this, and after another friend told a story about a guy hot knifing a spider's body, thinking it was hash, he said "Hey, here's $10 man, see that dead spider?"
"You fucking bastard" was my only reply...I smoked that fucking thing, and it sucked, but it could've been worse...tasted like a burning mushroom.

It made me kind of vibrate...could've been poison, could've been nausea, could've just been in my head, but regardless the feeling passed, and I got my smokes...

All in all, it was worth it.

By: JG

JG — Thu, 17 Aug 2006 21:06:34 +0000

I don’t know about the ECPA, but what about the DMCA? We as Web 2.0-enlightened people all believe that consumers are now content creators, right? So isn’t a query I’ve typed essentially a “performance”? And if not a query, then a whole series of queries? A whole series of queries is pretty substantive, as we have all found out recently. Don’t I, as the “performer” of those queries, have a right to control their re-publication, as per the DMCA?