May 3, 2006 at 10:00 AM ET
There's Bob Sullivan, the Red Tape Chronicles author. Then there's Bob Sullivan, who might be a bankrupt child molester with a brother who's a killer. One is flesh and blood, one is a computer creation. But in our digital age, who's to say which one is real? If perception takes on its own reality, certainly a computer creation can, too.
If you use the Internet today to conduct a background search on me, you might get the idea that I have been convicted of child molestation, and I have a close male relative who's been convicted of manslaughter.
Let me assure you, neither is true. But let me try to convince you that there is a crisis at hand.
Databases are spinning out of control. Our country is awash in data that is secretly collected, inaccurately transcribed, sloppily connected. As they say in the business, there's a lot of dirty data. But dirty data is being used to make important decisions that affect us. It decides who gets home loans and car loans, who gets credit cards, who gets insurance, even who goes to jail.
And sometimes, it's used to falsely taint people as potential child molesters and murderers.
Today, if you went to a number of prominent Web sites -- including Yahoo.com -- and decided you wanted to purchase a background check on me, you'd be led to a service named Intelius.com. The site's motto is "Building Trust." Until recently, it was "We know."
Intelius is one of hundreds of sites that offer consumers the ability to perform background checks on anyone. No license required. No permission slip required. Just type in a name and a state, and up comes a list of potential targets. Select a name from a list that looks about right, pay about $50, and off you go.
It is Intelius.com's computers that seem to think I might be a molester with a murderous relative.
Connection to dot-com mania
The site is run by data maven Naveen Jain, who rocketed to fame and fortune in the go-go 1990s as head of former dot-com darling Infospace.com. At the height of Internet mania, Jain was worth $8 billion, and Infospace worth more than Boeing Corp. Infospace stock took an historic tumble when the bubble burst, spurred in part by questions about the company's revenue projections. In 2002, Jain was replaced as CEO.
Soon after, Jain founded Intelius. The firm offers background checks and identity-theft protection services. Consumers can sign up and receive warnings any time there's a new account opened in their name, or there's a public record indicating an unusual event like a change of address. Jain says he has 3 million paying customers now. After receiving an invitation to Intelius' Bellevue, Wash., offices last year, I tried out the background service.
I was in for a big surprise.
'Child molestation 1'
Using Intelius is easy enough. You type in a name and a state; then you are presented with a list of potential "hits." There were a few dozen Robert Sullivans in Washington state; I picked the one who lived in my hometown, whose age matched mine, as a neighbor might do. And I agreed to pay the $50 fee.
The report I received was seven pages long.
Under a section titled "Criminal Check," two possible convictions were listed. One indicated it was for an unspecified offense. The other charge:
"Child molestation 1."
There's also a civil judgment listed under my name, a bankruptcy filing.
But that's not all. Under the section "Possible Relatives and Associates Report," Intelius lists the names and phone numbers of my parents, my sister and a mysterious "Shawn Sullivan." When I clicked on Shawn's name, I received additional details about him. Shawn is listed as having a possible conviction for involuntary manslaughter.
Again, none of these things are true about me.
To be fair, the Intelius report comes with qualifiers. The biggest one of all, in small print, says this: "You should not assume that this data provides a complete or accurate history of any person's criminal history." Newer versions of the report contain an more extensive disclaimer: "It is important to understand that public records are only as accurate as the agencies that input them. Please be sure to closely review the public information listed about the individual that you may be researching in the report."
Above my criminal report data, there's a warning which indicates there are no records of convictions against someone with an exact match of my name and birthday in Washington state. But since some criminal records aren't filed with birthdays, the report explains that it includes all convictions for any Robert Sullivan with my middle initial and a Washington residence.
The Robert Sullivan who apparently ended up on my background report was born in 1943, a fact listed next to the conviction record.
Hard to unring a bell
Still, is that enough to prevent any tainting of my name, any possible guilt by association? As privacy advocate Rob Douglas, who operates PrivacyToday.com is fond of saying, "It's very hard to unring a bell."
As for my new brother Shawn, his appearance on my report is a database curiosity. Shawn and I both lived in Columbia, Mo., at one point in our lives. His address was 211 Waugh St., mine 211 N. Ann Street. Because Intelius' matching technology only compares the city and the numeric part of the address to form an association, the computers assumed we lived together. Shawn and I are now digital brothers.
Again, there is the disclaimer that he is a "possible relative." But given that the rest of the data in that section of the report is accurate, a searcher would probably be led to believe Shawn and I have something to do with one another.
And remember, this searcher could be anyone: a neighbor, a friend, an enemy. Anyone can look you up, too. And these reports that are jumbled together to create a picture of me and you -- they cannot be fixed. I have asked Intelius to remove this wrong information from my report. The company won't. In fact, according to Jain, there is nothing wrong with my report.
"There are no mistakes," Jain says. "It's just data."
One might call them "neighborhood," as opposed to precise, searches. I call them "over-broad" searches.
Whatever the name, this, I would argue, is madness.
Helping parents keep kids safer
Jain forcefully defends his background reports. Parents who are hiring babysitters for their kids need every opportunity to unearth the possibility that the stranger they may be letting into their home is a molester. Intelius provides the broadest possible results, Jain said. If the results were narrower -- say, limited to exact name and birthday matches -- a molester might slip through the cracks. Jain offers the usual defense in these cases.
"If I can prevent just one child molester from getting at a kid, it's worth it," he said.
While that standard may be good enough for the average person, it's not good enough for companies or landlords running background checks on potential employees and renters. For those, the name and birthday must match exactly, Jain says, according to the tenets of the Fair Credit Reporting Act. That means professionals who use these reports for a living get a more limited report – and casual users who may not quite understand what they are looking at see all the additional potential sex offenses.
"I am convinced that what we are doing is the right thing to do help protect the families who use our service," Jain said. "You may question our judgment on erring on the side of the protecting the families, but we believe that, at this time, it is in the best interest of our customers."
One important note about such over-broad background searches – they tend to be heavy on sex offenses because there is so much data available on sex offenses thanks to Megan's Law, which requires sex offense registries. Background search companies that want their clients to feel like they are producing thorough reports use broad search queries to load their reports up with a few sex crime records. After all, when a clean report is returned, it looks like a waste of $50.
Accurate or not, the sex conviction will stay on my report. But after our talk, Jain said he's considering an option for consumers to add a note to reports like mine, advising anyone who pulls it up to pay special attention to the birthday, hair color, or other factors that would show the conviction wasn't really theirs.
"I think having something on top of the background check report about a person with an explanation from a person should be very effective in providing complete perspective," Jain said.
Again, I say, this is madness. Why should I have to do defend myself against Intelius' database? And the more important question: How many Inteliuses are out there? How many services am I supposed to use and research myself to make sure there are no errors? How many statements will I have to add arguing for my own innocence?
And then there's this question, familiar to any law student – how does someone prove they are not beating their wife or husband?
Other stories of data gone crazy
This is just one story of data gone crazy. There are many others. Famed cybersleuth Richard Smith once looked up his ChoicePoint report, and says the report suggested he was dead. But at least ChoicePoint makes its reports available to consumers for free, and ChoicePoint provides instructions on correcting mistaken reports. Others are not as approachable.
A consumer wrote to me recently after she had looked herself up on another backgrounding site and got a surprise like mine. She says her report listed erroneous convictions against her. When she looked up the site, there was no way to complain. No phone number, no e-mail address, no mailing address.
"What am I supposed to do?" she asked me. "I can't sleep at night."
Someone should give her an answer. While Congress has spent an interminable amount of time debating legislation designed to sprinkle regulations on the commercial data brokerage industry after last year's embarrassing data leaks, the world has continued collecting billions more bits of data about us.
And no one is really thinking about what will happen to all that information, all that dirty data. It's time we did.