home

What Google Searches and Data Mining Mean for You

Siva Vaidhyanathan of Sivacracy is guest posting at Altercation today. One of the issues the NYU prof and media scholar tackles today is feds demanding records of google searches and data-mining.

As many of us have been saying for years, there is real danger in Google (and Microsoft, Yahoo, AOL, Amazon.com, etc.) collecting dossiers on our Internet habits. While these companies might annoy us with targeted ads and please us with customized service, they collect all sorts of information that might or might nor accurately reflect our relative danger to society.

The danger comes when the federal government demands records of our Internet usage and then runs the data through badly-designed "data mining" algorithms. In the specific case of Google, the Feds are not seeking any individual or identifiable information. But spikes and anomalies in the data could send them scurrying for more specific information. If it sees a high incidence of searches for "Anthrax" (the microbe, not the band) or "Jihad" over a set of days, the FBI could request deeper data. Then you could be in trouble.

For instance, you might want to read all you can about Al Qaeda before you take the Foreign Service Exam. Your innocent searches could be misinterpreted as solidarity instead of curiosity. All of a sudden, you find yourself a "false positive." Worse, such a foolish process can't help but yield "false negatives," letting really bad people slip through while we place way too much faith in the power of raw data and fast computers.

As to the import, Siva says it's very important - and ridiculous. He explains why. He then adds that the reason this is possible in the first instance is because Google is allowed to keep our records.

....the real problem here that Google collects and keeps all our data in the first place. We must demand that Google purge its servers of our personal data. If it won't we should demand that Congress act to protect us. I know, the thought of Congress actually working to protect Americans is painfully out of date. But hey, I can't give up hope.

< Poll: More Americans Favor Impeaching Bush | High Court to Hear Death Case Challenging Execution Procedures >
  • The Online Magazine with Liberal coverage of crime-related political and injustice news

  • Contribute To TalkLeft


  • Display: Sort:
    Re: What Google Searches and Data Mining Mean for (none / 0) (#1)
    by squeaky on Wed Jan 25, 2006 at 01:56:20 PM EST
    Very well put. Even the wingnuts would have to agree. Imaging your google searches are flagged by a computer because you are writing your PHD on terrorism with the hope of going into government foreign relations work as a Mid -East diplomat. You are put on a no fly list. You get falsely arrested and "the wall" between fed and local agencies has been torn down. All of a sudden you are in the system with tons of supporting misinformation generated about you. Your case gets a gag order. Soon your innocence is doubted by normal people. Every quirk in your personality suggests that the charges are true. Normal people will believe you guilty because it is too scary for them to imagine the system could be so faulty that the same thing could happen to them. A fear defense mechanism. Next thing you know you are in gitmo with a bag tied over your head. I agree that information should not be stored in the first place. Sadly there is too much money to be made for this practice to stop. So the idea of probable cause has to come into play and bulk info does not belong in the hands of those who have no reason to have it.

    Re: What Google Searches and Data Mining Mean for (none / 0) (#2)
    by jimakaPPJ on Wed Jan 25, 2006 at 01:57:51 PM EST
    Okay, we agree. Now, where is the outrage from the Demo Congress people?

    My understanding is that Google keeps track of people by using cookies and IP addresses. I not surprised that google keeps this information. I am surprised they do not anonymize it. It is pretty easy, to take a string like a cookie and apply a one way mathematical operation to it to map it a new string that is unique for that cookie but that cannot be mapped back to that cookie. Why do they keep this personal tracking non-anonymized? That is disturbing and is ripe for exploitation.

    Re: What Google Searches and Data Mining Mean for (none / 0) (#4)
    by The Heretik on Wed Jan 25, 2006 at 02:32:56 PM EST
    Just don't google something like "rent boy" and your problems will be far less.

    Re: What Google Searches and Data Mining Mean for (none / 0) (#5)
    by Slado on Wed Jan 25, 2006 at 02:49:51 PM EST
    This whole issue is being blown out of proportion. The department of Justice involved in a lawsuit with the ACLU about child porn requested raw data from all the major search engines (all complied but Google) and the data did not include personal information. What's the problem?

    Re: What Google Searches and Data Mining Mean for (none / 0) (#6)
    by Slado on Wed Jan 25, 2006 at 02:52:57 PM EST
    Also shouldn't it really be a concern that in the same breath they will give special software to the Chinese so they can keep their citizens from veiwing everything the internet has to offer, including porn? That seems more serious then the hypothetical worries of the blogoshpere.

    Some interesting info could be gleened from the searches, even without the identifying info. Regarding his UCLA claims, could a copyright lawyer verify them? If a lecture is recorded on tape, does it not then become copyrightable due to the fixed nature of the tape medium? And, obviously, the students who'd do this aren't doing it for the $2 an hour, they're doing it to keep profs honest.

    Re: What Google Searches and Data Mining Mean for (none / 0) (#8)
    by Sailor on Wed Jan 25, 2006 at 03:21:03 PM EST
    Okay, we agree. Now, where is the outrage from the Demo Congress people?
    Jim, do you EVER comment on topic? With your rhetorical statement, then the only answer is 'since repubs control congress where is their outrage?' And the thread is hijacked. If you care to respond I'd be happy to answer on an open thread. Back on topic: After my courses in internet security, and with the advice of the instructor, I established a policy of all original server logs being destroyed every 2 weeks and only summations of traffic are kept. Google, and every site should be mandated to the same.

    Re: What Google Searches and Data Mining Mean for (none / 0) (#9)
    by Johnny on Wed Jan 25, 2006 at 04:23:17 PM EST
    Slado, your obeisance to the admin is slightly disturbing. The thing I find most disturbing about the gov't request is their demand of 1 million random websites.

    Re: What Google Searches and Data Mining Mean for (none / 0) (#10)
    by jimakaPPJ on Wed Jan 25, 2006 at 04:38:15 PM EST
    sailor - You are hard to please. I agree. Now. I think you could agree that this isn't going to be pushed by the Repubs. So why beat them? It will do zero good. So my question is on target. Where are the Demos? Surely if they can worry about the wire taps on terrorists, they can worry about this. You know what's funny? The Demos can win on this issue, but will lose on the other. Which one will they choose?

    Re: What Google Searches and Data Mining Mean for (none / 0) (#11)
    by Sailor on Wed Jan 25, 2006 at 04:57:55 PM EST
    So we all agree that the gov't has no right to this data?

    Re: What Google Searches and Data Mining Mean for (none / 0) (#12)
    by jondee on Wed Jan 25, 2006 at 05:02:25 PM EST
    Am I still safe going to Hot Circus Dwarves?

    Re: What Google Searches and Data Mining Mean for (none / 0) (#13)
    by Sailor on Wed Jan 25, 2006 at 05:39:07 PM EST
    Jondee, as your personal NSA rep, I would say going to dwarves is OK, googling them isn't;-) BTW, we have our i on you.

    I googled Rent Boy and all I got were photos of Dubya and Jeffy Gannon. Google ROCKS! They are standing up to the fascists. Gonzo is using an UNCONSTITUTIONAL law to demand this data. Yahoo and AOL can GO TO HELL.

    Re: What Google Searches and Data Mining Mean for (none / 0) (#15)
    by BigTex on Thu Jan 26, 2006 at 09:40:51 AM EST
    While I agree the government shold not have access to the data, the data should be stored to block out access to child porn. I'm not suggesting that the users who look at the trash be turned over, but rather that the web pages themselves be turned over. Google has demonstrated the capability to shut out information. They are "sanatizing" the internet to make China's government happy. The same technology should be used to sanatize out child porn in the US. The individual users don't have a right to privacy, but the government shouldn't be allowed to go on a fishing expidition either. But web pages aren't people who have 4th Amendment Rights (what little are left.)