FBI Data Mining Program Raises Eyebrows in Congress  

The Blotter | June 13, 2007

By Justin Rood

Lawmakers are questioning whether a proposed FBI anti-terrorist program is worth the price, both in taxpayer dollars and the possible loss of Americans' privacy.

The National Security Analysis Center (NSAC) would bring together nearly 1.5 billion records created or collected by the FBI and other government agencies, a figure the FBI expects to quadruple in coming years, according to an unclassified FBI budget document obtained by the Blotter on ABCNews.com.

Those numbers alone raised concerns from two congressmen, Reps. Brad Miller, D-Calif., and James Sensenbrenner, Jr., R-Wisc., the chair and ranking member of the oversight panel of the House Science and Technology Committee.

The FBI has a track record of improperly -- even illegally -- gathering personal information on Americans, most recently through the widespread abuse of so-called National Security Letters, the two men noted in a letter to Congress' investigative body, the Government Accountability Office.

Miller and Sensenbrenner asked GAO to determine whether the NSAC will include records on U.S. citizens, and whether there are protections in place to make sure all the data in the program was legally collected.

Of further concern to the two congressmen are the FBI's stated hopes to "pro-actively" mine the data to find terrorists using "predictive" analysis, according to its budget request, an unproven method according to experts and even the U.S. intelligence chief's office.

In theory, predictive analysis involves mapping a known pattern of terrorist behavior -- for instance, the sequence and timing of such mundane activities as bank transactions and travel purchases -- against a massive collection of such records like the NSAC databases. If an individual's actions match the pattern, they can be considered a suspect, even if they have no known ties to any suspected terrorists or known terrorist groups.

Such a method would help identify "sleeper cells," the FBI claims in its request -- secret groups of terrorists living innocuously within the United States, waiting for a signal from a terrorist group leader to assemble and strike.

But to date the approach has not proven workable. So far, terrorism researchers "cannot readily distinguish the absolute scale of normal behaviors" for terrorists or ordinary Americans, conceded a 2006 document from the Office of the Director of National Intelligence and obtained by National Journal magazine. In other words, no one can figure out how terrorists act differently from normal Americans.

"We had no idea how on God's earth you would characterize and capture normal behavior," a former researcher for the ill-fated Total Information Awareness (TIA) program told the magazine last October.

TIA, the government's first attempt at anti-terrorism data mining on a massive scale, had its funding stripped by Congress over widespread concerns it would violate privacy laws. The National Security Agency -- arguably a more tech-savvy outfit than the FBI, whose computer woes are legendary -- continues to pour millions into data mining research.

The FBI has requested $12 million for its NSAC project. That amount would pay for 90,000 square feet of space and an additional 53 employees, according to its budget request. Whether Congress will approve the funds has yet to be determined.

The bureau did not respond to a request for comment.