Garbage In, Unnecessary Arrests Follow

(Originally posted on Daily Kos.)

We all know the government has a love affair with data and data mining, and while Congress tried to kill the Total Information Awareness program, it never quite went away. One of the key arguments against data mining is that bad data in the system would increase errors exponentially, creating huge privacy and civil liberties problems for innocent people and wasting resources by generating a huge number of false leads (and you don't have to trust us, the National Academy of Sciences said so). Well an article that ran in USA Today lastWednesday gave us yet another reason to worry about the quality of data in government databases.

The article tells the story of Ann Wright, a former U.S. State Department official who became active in the anti-war movement after the invasion of Iraq. Wright has been repeatedly stopped at the Canadian border by immigration officials because our government gives them direct access to a database that includes her history of misdemeanor arrests for civil disobedience. Wright's story is certainly one worth telling, and highlights a growing problem of our government's sharing sensitive personal information about Americans with foreign countries, often with serious consequences.

BUT, there's an interesting fact buried in the article that we thought was worthy of some attention. It's not just that our government is sharing derogatory personal information about Americans with foreign countries, it's that much of the information is wrong!

The database the article discusses is managed by the FBI's National Crime Information Center (NCIC), the government's primary computer system for sharing criminal justice information with law enforcement agencies throughout the country, and now beyond, including information about fugitives from justice, terrorists, and missing persons. NCIC process an average of 7 million requests for information a day. The article quotes Roy Weise, a "senior adviser to the FBI's Criminal Justice Information Services Division," which manages NCIC, as saying that since 9/11, the FBI has felt a greater need to share this information. But the bombshell comes next:

"About half of the arrest records in the system have not been updated to reflect convictions, dismissals or acquittals, Weise said, adding that local law enforcement agencies are responsible for giving the FBI updated information."

This is not a new problem; it's only new that the FBI is admitting how bad it is.

In 2005, the Migration Policy Institute conducted a survey that found an error rate of 42 percent in NCIC "hits" between 2002 and 2004. You might wonder how NCIC could be so error-prone, given that the Privacy Act requires agencies to "maintain all records which are used… in making any determination about any individual with such accuracy, relevance, timeliness, and completeness as is reasonably necessary to assure fairness to the individual in the determination." Turns out, the FBI exempted NCIC from this Privacy Act requirement in 2003.

With a database as important to the security of Americans as NCIC in such a sorry state, you can now understand why even less regulated, but equally important databases like the Terrorist Screening Center watch list are such a mess.

We need to start scrutinizing the processes our government uses in collecting, analyzing and sharing data about Americans that can have serious consequences to their rights and privacy. Congress should be mandating that agencies follow clear and strict procedures for placing personally identifying information about Americans in these systems, and should conduct stringent oversight. It's time to put watch lists on our to-do lists.