GAO finds widespread fed data mining

According to a new study, data mining is more pervasive in the federal government than many citizens may realize, a finding that government auditors suggest may exacerbate privacy concerns.

A General Accounting Office report released this week reveals 52 federal agencies that use or plan to use data mining, in most cases for reasons other than searching for hidden terrorist activities. At least 122 of the data mining projects reported in the study include the use of personally identifying information, some of it held in private-sector databases.

For its study, GAO looked at data mining that relies on statistical and modeling techniques to help analysts discover hidden patterns and relationships and to make predictions based on those discoveries.

As GAO noted in its study of 128 federal departments and agencies, data mining challenges the concept of privacy protection afforded to individuals "through the inherent inefficiency of government agencies analyzing paper, rather than aggregated, computer records."

Among the 52 agencies, 199 separate data mining projects are underway or are planned. Service improvement accounts for the largest number of projects, and analyzing intelligence accounts for the smallest number, the study found.

Not surprisingly, the Defense Department accounts for the largest number of data mining projects. The Education Department reports the greatest portion of data mining activities for discovering fraud, waste and abuse. NASA is the source of the most numerous data mining efforts at analyzing scientific information.

Auditors found the most common use of data mining among all agencies is for detecting patterns of criminal activity. GAO conducted the study between May 2003 and April 2004 at the request of Sen. Daniel Akaka (D-Hawaii), ranking minority member of the Senate Subcommittee on Financial Management, the Budget and International Security.