Databases

In reading my morning paper, I found this AP report:

STAMFORD, Conn. - Jessica Smith thought she was a shoo-in for a cashier's job at an Office Depot in Minnesota this past summer.

The store manager was encouraging, saying he just needed to run a criminal background check.

But a week later, Smith received a rejection letter that cited a lengthy rap sheet, including drug convictions in Washington.

"I have no record," Smith, 19, said as she flipped through court documents. "They all say felony and guilty. I've never even been to Washington."

Smith, who fought for six weeks to clear her name before eventually landing the job, was a casualty of one of the latest trends in business hiring. Companies increasingly rely on pre-employment background checks to ease security concerns and protect against costly lawsuits.

I'm an unabashed civil libertarian. How many cases like this occur where an individual is actually damaged and doesn't pursue it, or where the damage is not revealed? Or perhaps you pop up on a screeen and get an FBI tail (with a record)?

The proliferation of databases, private and governmental, represents and incredible threat to the right to privacy. That's really not news. And those who argue that giving up some privacy for protection would disagree with my focus on privacy.

But let me add this.

I've worked on databases all my life. Most recently, I've been working on a campaign database for a candidate for which I'm volunteering. And like past experiences, the errors in data input and maintenance are enormous. Even after many hours of "clean-up", I wouldn't be surprised to find a data error rate of 10%. As the old saying goes, "shit in, shit out", databases are only as good as the input and the maintanence. And make no mistake, the larger the database, the more likely the errors.

Creating and maintaining databases is a very laborious and therefore expensive endeavor. A governmental database, usually mined and compiled from many sources, becomes a herculean task perhaps involving thousands of individuals (or varying skill sets) to keep accurate. And no one who works with this information will guarantee it's accuracy. So anytime anyone tries to sell the "efficiency" of using information to catch bad guys, or screen for terriorist, or for any other purposes of security, just remember that for every possible catch, there may be hundreds if not thousands of false hits, resulting in enormous inconvience if not damage to those involved.

Databases have an incredibly weak link.

People.

Computers are stupid and depend on people for efficiency and accuracy.