Senate hearing will examine data mining

In his first hearing Wednesday as the chairman of the Senate Judiciary Committee, Democratic Sen. Patrick Leahy of Vermont plans to examine federal “data-mining” programs, the computerized hunt for terrorists that can implicate innocent people.

Consider the case of American Airlines pilot Kieran O’Dwyer of Pittsboro, N.C.

O’Dwyer said Tuesday that U.S. Customs agents detained him for 90 minutes in 2003 when he got off an international flight in New York, telling him his name matched one on a government terrorist watch list.

Over the next 22 months or so, O’Dwyer said, he was temporarily detained 70 to 80 times by authorities who apparently were worried that he was a fugitive member of the Irish Republican Army.

It’s not clear how O’Dwyer came under suspicion or how his name wound up on a terrorist watch list, but critics charge that such miscues can occur during the data-mining process, in which computers analyze multiple databases in search of suspicious patterns.

Amy Kudwa, a TSA spokeswoman, said she couldn’t comment on O’Dwyer’s circumstance, but that an average of 1,500 airline travelers applied each week for redress on the grounds that they’d been mistakenly included on terrorist watch lists. She said 33,000 had applied as of last April.

In a speech last month, Homeland Security Secretary Michael Chertoff said his agency was working to put a system of “one-stop redress” in place in 2007.

Leahy, a longtime congressional champion on privacy issues, plans to make “a signature issue” of protecting civil rights in the face of a “proliferation of government databanks and data mining” in the war on terrorism, said his chief spokesman, David Carle.

“He believes Congress is way overdue in taking stock of the surge in data mining by the government,” Carle said.

Carle said the inquiry had gained import because of powerful new technologies, the outsourcing of data mining to private firms and “the Bush administration’s lack of cooperation” with Congress’ attempts to police these surveillance programs.

For years now, the Bush administration has invested heavily in data mining, viewing it as a valuable intelligence tool that can alert U.S. authorities to terrorist plots in their early stages. The government is reported to have spent tens of millions on such surveillance systems.

Among witnesses summoned to Wednesday’s hearing is Jim Harper of the Cato Institute, a libertarian research center. Harper co-authored a paper last year that concluded data mining is a poor use of tax dollars, won’t identify terrorists and will lead to “false positives” implicating innocent people.

In the paper, Harper and IBM engineer Jeff Jonas charge that “the one thing predictable about predictive data mining for terrorism is that it would be consistently wrong.”

In testimony prepared for delivery Wednesday, Harper says data mining uses “massive amounts of data about Americans’ lifestyles, purchases, communications, travels and many other facets of their lives. . . . This raises a variety of privacy concerns.”

Coinciding with the hearing, Democratic Sen. Russ Feingold of Wisconsin plans to reintroduce a bill, co-sponsored by Leahy, that would require federal agencies to disclose all data-mining activities to Congress and the public.

In a 2004 report, the General Accountability Office identified at least 10 data-mining programs being used in the hunt for terrorists, and several others have emerged publicly since then.