Add new comment

AOL's ill-considered decision to release 19 million search queries conducted by over 650,000 users from March through May of this year has broken out of the blogosphere into mainstream media outlets like CNN, the BBC, MarketWatch and WebProNews.

As far as I can tell, Greg Linden was the first person to point to the data set on AOL's research site, back on August 4th, and the comments on his post highlight the potential value of this attention data and AOL's slapdash approach to anonymization. For example, "Reto" commented:

I've now had a chance to spend 5mins browsing the data myself. There's a couple hundred examples of people pasting a phishing email into the search box, each email begins: 'Dear [user's *full* name]'.

I picked a name at random and was quickly able to see this guy lives in Ohio but is moving to Georgia, he drives a Chevy van (which he's looking to 'pimp'), he may have stomach cancer, he desperately wants to win the lottery -- and is considering enlarging his...

Ahh, in any case I know his full name and where he lives, combined with *very* personal details.

Adam D'Angelo at CalTech was also one of the first to assess and discuss the data set:

User 491577 searches for "florida cna pca lakeland tampa", "emt school training florida", "low calorie meals", "infant seat", and "fisher price roller blades". Among user 39509's hundreds of searches are: "ford 352", "oklahoma disciplined pastors", "oklahoma disciplined doctors", "home loans", and some other personally identifying and illegal stuff I'm going to leave out of here. Among user 545605's searches are "shore hills park mays landing nj", "frank william sindoni md", "ceramic ashtrays", "transfer money to china", and "capital gains on sale of house". Compared to some of the data, these examples are on the safe side. I'm leaving out the worst of it - searches for names of specific people, addresses, telephone numbers, illegal drugs, and more. There is no question that law enforcement, employers, or friends could figure out who some of these people are.

This is obviously bad news for AOL and worse news for AOL users whose privacy has been violated. But hopefully this episode will contribute to a broader understanding of attention data and encourage people to treat their personal data as a valuable resource.

tags: attention attentiontrust attention+trust attention+data attention+economy aol aol+research greg+linden adam+d'angelo