Lawyers call for data mining guidelines

There must be a framework for using data-mining technology, legal experts tell a House committee.

Data mining can be an important tool as the government collects vast quantities of intelligence for homeland security, but there needs to be guidelines and policies, experts told the House Permanent Select Intelligence Committee Thursday.

Technology makes possible many analyses functions, but agencies should not be allowed to simply use them however they see fit, said Judith Miller, a partner at the law firm of Williams and Connelly LLC, and former general counsel at the Defense Department.

"Although these technologies are probably important for our national security to pursue, we have not been going about it so far in the right way," said Miller, a member of the Markle Foundation Task Force.

Rep. Jane Harman (D-Calif.), ranking member on the committee, voiced concern that the Bush administration has "not yet adequately developed a proper policy framework for data mining." Right now the government is still functioning on an "emergency" approach with homeland security rather than developing long-term policies, she said.

That type of framework is necessary for the executive branch — and Congress — as data-mining issues increasingly appear in federal agencies, Miller argued. It should include a governmentwide review process, consistent implementation guidelines, and oversight from Congress and the administration through a clearly established audit trail, she said.

The most controversial data-mining project under the homeland security mission so far has been the Total Information Awareness (TIA) program at the Defense Advanced Research Projects Agency. TIA would search through open source, third party information — a type of search that is not prohibited by any regulation or legislation, pointed out William Barr, executive vice president and general counsel at Verizon, and former U.S. attorney general.

However, given the civil liberties concerns around projects such as TIA, technology will not be the answer, said Philip Heymann, a professor at Harvard University School of Law and former deputy attorney general. Significant legal, ethical and cultural decisions will need to be made every time agencies develop a new homeland security data mining program, and "I think only human beings can do it," he said.

NEXT STORY: Courts can cover IT budget cuts