<< Never say never? >>
Of course. He's just being realistic. All it would ultimately take is one Googler to make a wrong decision and they could be in the same situation. Unless he personally approved every decision in the company, he wouldn't be able to guarantee this wouldn't happen at Google.
(Having said that, I don't believe that the decision to release AOL's data was down to just one person – it's unbelievable that nobody on the team noticed the uproar it could cause. Of course, it's possible that they knew and just thought there's no such thing as bad publicity...) |
I hope it takes more than 1 Googler... I bet the AOL decision to release the data bubbled up through at least some hierarchies/ several people (yes, we know it didn't reach the highest hierarchies, but still), and that it takes a certain company philosophy not putting privacy as very high goal to make this go through. If you drill your company to respect privacy you decrease the likelihood of this happening, and if you don't, you increase it. I don't know how well Google fares on this gradient.... and what Google does to prevent this. |
This is unrelated to the topic but it concerns google and I figured that you might
be interested in it (I have been a long time RSS reader :)).
Blogger has a few security problems and one of them has been exploited by either a smart spammer or a "funny guy" to play with Abrupto. Abrupto is a former technorati 100, #1 for portuguese blogs and is written by JPP a well know portuguese politician/opinion-maker an former VP of the European Parlament.
The attacks have been going on for almost 3 weeks, jpp has written 17 posts about this (he numbers them :)), this is the last http://abrupto.blogspot.com/2006_08_01_abrupto_archive.html#115514771603966338
hasreceived some emails from blogger and 3 or more times he was promised that it would end. It has not, last night the pirate's page appeared once again, in fact the post is still at http://abrupto.blogspot.com/2006/07/this-is-my-first-blog.html
This mike dude has been playing innocent (what happened to my blog?), changed names, changed locations and he also serves a few ads on the side bar.
Regardless of their promises blogger hasn't been able to put an end to this and everytime they assured that it was over the issue reappeared shortly.
If you would like more info on this or portuguese ios not your strong you'll find my contacts at my place, mouronacosta-wordpress-com |
Or compare what the companies did when they were subpoena'd for search logs a while ago:
- Google tried to resist this
- Yahoo resisted it in parts
- AOL and Microsoft fully complied
It's not even a moral thing, necessarily: if an IT company would be completely selfish thinking just of increasing revenue, they understand that privacy is connected to gaining user trust, and user trust is connected to future revenues. |
He means "Google is so secrecy-oriented that'd we'd never *dream* of providing complete reseach datasets, so you don't have to worry that we would goof like that!" :-)
|
This transcript of Eric's Q&A gives a better look on Eric's emphasis:
<<This [AOL data leak] is obviously a terrible thing. The data was not anonymized and it was a mistake. If Google were to make this mistake, it would be a terrible thing. They have lots and lots of systems to prevent this from happening at Google. They don't share everything in Google with everyone in Google. He describes a case where the government gave Google a subpoena that was over-broad, and they fought it in court. They take is so seriously that they fight it in court.>>
http://www.seroundtable.com/archives/004343.html |
Speaking of how Google fought off the DoJ when it requested data from Google ... I find it very interesting that the data AOL let out is the same data the DoJ wanted, and since Google powers AOL search ... and in the guise of a research area, and then a "Wow, my bad," all the CEO's have plausible deniability built-in.
How very convenient. |
"Google recognized those possibilities earlier this year, when the U.S. Department of Justice demanded that they turn over two month’s worth of users’ search queries. Google challenged the subpoena, and a federal judge sided with the technology giant. Upon learning of the judge’s ruling, Google declared the win “a clear victory for our users and for our company.” But AOL and several other search companies complied. So if you use anything but Google—and are prone to ego searches, looking up directions, and finding restaurants near your home—there’s a good chance that the Justice Department knows exactly what you searched for in June and July of last year. "
http://www.taemag.com/issues/articleID.19316/article_detail.asp |
AOL established a research team which then put up a wiki to share data and their findings. Now that they've been crucified for exposing too much data, it's very unlikely that sharing data for research will be easy in the future.
Google's data sharing is more controlled, but they are doing some.
Google is releasing a terabyte of 5-word n-grams to LDC, which individuals cannot really access. It costs $25k to join LDC.
The n-gram data isn't session-connected, so it's probably not gonna have the same impact, but it certainly COULD contain names / SSN / credit cards, though it's much less likely since n-grams are likely filtered by occurrence.
Somehow, I think people would be better off if search session data was more explicitly public. The data is out there, but as it stands, only small numbers of people are able to access – those at search engines, big websites, ISPs, network access providers, etc. and those who can afford to buy it from those sites. |
John, how I understood it is that Google's upcoming N-gram data will be compiled by analysis of the public web pages (not private web searches).
> The data is out there, but as it stands, only
> small numbers of people are able to access
Thanks to AOL, that problem seems to be solved :) |
This was also posted here
http://blogoscoped.com/forum/61339.html#id61339
;-) |
