Hacker News Reader

Comment by nyellin | Hacker News Reader

Google Search has bigger privacy issues. IP Addresses are anonymized after 9 months, but your search history is stored forever in anonymized form. If you have ever searched for yourself on Google, or for a unique set of information that can identify you as the origin of your searches, then you have to trust Google's security team and future lawmakers to keep your information private.

Yes, most people don't care, and I'm not any better. I don't have anything incriminating or too embarrassing in my search history, so I use Google and grit my teeth whenever privacy comes up. I tried DuckDuckGo for a week, but I eventually fell back to Google.

Sources and relevant links:

http://www.google.com/privacy/faq.html

http://en.wikipedia.org/wiki/AOL_search_data_scandal

View on HN

I agree that I could care less about all of this stuff, it's the search data that matters. Imagine taking a look at a month of someone's search data, it would be revealing (it might even be revealing to look back over years of your own search history).

But as for the question of forever anonymization I think you have misunderstood:

The data will not be traceable back to a single computer or user. The link between searches is the IP address, and it would be pretty hard to identify a single IP address without the final two octets (ie 255.255.255.255 -> 255.255). It definitely wouldn't be impossible (nothing is), but I would bet it would be within the level of difficulty that many other more privacy-violating things would.

Furthermore, my ISP changes my IP address very frequently (my guess would be about every 48 hours), and even if it was static, I would not have had the same IP for a very long time (moved frequently, etc).

The cookie is the only link between the IP addresses, once you've ambiguated the IP's it's pretty hard to connect the numerous IPs one person would have to people.

The sources you submitted were not very explanatory and have nothing to do with Google, respectively.

http://www.wired.com/threatlevel/2007/03/google_to_anony/

The other data stored with your searches could be way easier to track you from than your IP address: http://panopticlick.eff.org/

It would (theoretically) be easy to determine which searches are all by the same person using data like that, and then you just have to find one of those searches that reveals their identity.

You are mostly correct. According to an Ars Technica article from last year [1]:

* Google anonymizes the last octect of your IP address after 9 months. (So there are 254 possible IP addresses for each search.)

* The search cookie is kept intact for 36 months (3 years).

When I posted before, I incorrectly remembered reading that the search cookie is stored forever. Thanks for catching my mistake.

Edit: Regarding the sources, I think that Google's FAQ deliberately skimps on the details. I posted the AOL link because it is a good example of what can happen when search data is leaked.

[1] http://arstechnica.com/tech-policy/news/2010/03/google-keeps...

am I missing something with that 36 months? from the ars article:

"After 18 months, Google anonymizes the unique cookie data stored in these logs."

a 2.5 year old searchengineland article also says they use a hash function using keys that are discarded every day, which means that search streams will be 24 hours long apiece:

http://searchengineland.com/anonymizing-googles-server-log-d...

Urgh. I was severely sleep deprived last night and I am starting to feel that I shouldn't have posted at all. You are correct about 18 months.

> Furthermore, my ISP changes my IP address very frequently (my guess would be about every 48 hours), and even if it was static, I would not have had the same IP for a very long time (moved frequently, etc).

This isn't true for everyone. I have cable internet from Time Warner, and my theoretically "dynamic" address hasn't changed in at least a year.

You will be truly ANON, if you Install TOR and set preferences to delete all the cookies with each browser instance.

Imagine taking a look at a month of someone's search data, it would be revealing

Please explain to me how would any random person get that? Actually how would anyone get that data if you have not already done something which throws suspicion on you.

> ...and future lawmakers...

I figure in the far future some giant AI will probably be reading the data, for anthropological or historical reasons. Reconstructing all our lives as well as it can to understand us better, or maybe out of boredom. So whenever I'm going to do some really off-the-wall searches, just to see what there is out there on the web, or because of some strange conversation at the pub, I search first for something like "these next few aren't really about me".

Wonder if that might help at a trial too.

What about bing? If they have better privacy, then they can use it as a selling point to capture market share. A quick google search on google vs bing privacy policy seems to favor bing. I have the disconnect chrome extension to block google tracking.

I have Disconnect and there will still a lot of categories. They seem to direct you to the IBA extension to permanently opt out. There's another that supposedly works on more sites, but last time I tried it, it and Disconnect got into a race and crashed Chrome immediately, even after restarting it several times.

IBA Opt-Out: https://chrome.google.com/extensions/detail/gbiekjoijknlhijd...

Keep My Opt-Outs: https://chrome.google.com/extensions/detail/hhnjdplhmcnkieca...

Do I need both?

Reply on news.ycombinator.com