[disturbing]AOL's data sketch sometimes scary picture of personalities searching Net

Updated 8/8/2006 10:05 PM ET
By Kate Patterson, USA TODAY
America Online chief Jonathan Miller apologized for the incident.
AOL this week accidentally gave us a glimpse of what it's like to read minds, and now we have to figure out if we want to.

There's a dilemma here dramatized in a 1961 Twilight Zone.

A bank clerk — played by future Bewitched husband Dick York — suddenly gets the power to read minds. He thinks this is pretty cool, as most of us would. He finds out that colleague Miss Turner has the hots for him. He learns that his boss is having an affair.

And then it turns bad. Like, he finds out that a trusted old employee named Smithers is making plans to rob the bank and escape to Bermuda. York's character accuses Smithers — and then it's revealed that Smithers had for years fantasized about the plot but never intended to do it. York gets fired, and when he finally loses his ability to read minds, he is very relieved.

So this week AOL is getting hammered because a staffer, without authorization, posted on the Internet the millions of search words typed in by 658,000 users over a three-month period from March to May. The posting was supposed to be for researchers, who would study the patterns to try to create better search tools. But until AOL took it down, with deep apologies from CEO Jonathan Miller, anyone could download the data. Chunks of it are circulating on the Internet.

The data are anonymous: Each user's name is replaced by a number. But you can learn a lot about some of these people by the terms they type in.

User 1515830 is apparently an overweight woman with little willpower. She searches for the number of calories in certain foods but later searches for "baked macaroni and cheese with sour cream." Tragically troubled, she searches terms having to do with incest, depression, psychotic drugs and "I hate men." She also seems to be a teacher ("teaching positions in Denver Colorado"), probably lives in Ohio ("divorce laws in Ohio") and is shopping for curtains.

This is what's so fascinating about AOL's gaffe. AOL has made one of the secrets of search very real: It is the closest thing to the content of people's minds.

Never before has this kind of data been widely available. Yahoo and MSN gave search data to the government (and Google fought against that order), but the data stayed inside federal agencies.

Thanks to AOL, though, now we all know: Nothing ever invented — not the diary, not reality TV, not MySpace pages — gives us such a raw imprint of our thoughts, desires, needs and impulses as search.

You need help with your sex life, you type it into a search engine. You forget the lyrics to Bennie and the Jets, or never understood them in the first place, you type it there. Thinking about a vacation, worried about excess ear hair — for Internet users, if it comes to mind, you search for it.

We're candid when we search because it seems anonymous and private. "Most people assume the Internet is a more private medium than it is," says Greg Lastowka, a Rutgers law professor specializing in the Internet.

As AOL has shown, search records are, in fact, like a diary. The intention is that they remain private, but they can wind up in someone else's hands.

Reading people's minds through search data is powerful stuff. You can bet that Internet marketing firms are downloading the AOL data and analyzing it, industry experts say. The firms might be able to find general patterns of searches that often lead to expensive transactions. That would help them recognize when someone starts searching for words that often lead to, say, inquiries about purchasing a sports car. Then maybe Porsche ads could be targeted at those early queries, getting to a searcher before BMW.

Now that the AOL information is out and people can hammer at it, it might not be long before hackers start attaching real names to the numbers. "With enough puzzle pieces, in this case searches, it's very possible to figure out who someone is," says Jim Harper of the Cato Institute.

And how might certain people use that information? Blackmail? Public ridicule?

Meanwhile, voyeurs are digging through just for fun. It's wild to look inside other people's heads, imagining what's going on in their lives. News.com posted some of the more sordid search strings.

User 9486162 seems to be out of a job (he searches for "employment needed Louisville ky"), hooked on poker ("hold'em poker school") and considering suicide ("how to kill oneself by natural gas").

User 2708 seems to be a woman bent on payback for a breakup. She searches for "I hate my ex boyfriend," "how to humiliate someone," "free angry stuff to send to an ex lover" and "makehimpay.net" — which, much to any single guy's dismay, actually exists.

One of the most widely discussed search strings among bloggers belongs to user 17556639. He searched over and over for terms including "how to kill your wife," "murder photo" and "decapitated photo" — pausing in the middle of that string to search for "steak and cheese." Perhaps he got hungry.

Now what do we do with this new power? We can read these people's minds. We can assume someone is thinking of suicide, someone wants the privates of an ex-lover on a platter, and someone might want to kill his wife.

We could be like Dick York and send the police after the potential wife killer. But thinking of a crime doesn't mean he's going to do it. He could be a mystery writer doing research. He could be fantasizing after a nasty fight, and the couple will make up later.

But if he really is plotting to kill his wife, should search records be used to prevent it? Tough question.

These issues "will get bigger with more data collection," Harper says. More and more, the search industry will be able to read our minds. And there's no guarantee that information will remain safely tucked in some vault.

Do we like that? Welcome to the Twilight Zone.

__________________________-

I have a healthy distrust of Google, but AOL and Yahoo are in the same boat. The Census caught onto data miners and rearranged the data it releases so it wouldn't be so obvious. Amazon has its own profiling program, so this is just the start.

 

The cookies on your computer expire then but Google and the like probably have their own personal records. One of the electronic watchdog groups warned people to delete their Google cookies before logging onto Gmail for this reason.

But who really knows how aggressive these search companies and their free emails really are. For now, it's something to think about.