Digital Spelunking
Data mining failure
A new study from the National Research Council suggests that data mining-high-tech sifting through the digital leavings of our lives-won't catch the next Osama bin Laden.
The idea of scanning Facebook pages, online chats, blogs, Internet-based phones, and GPS tracking data for "unusual" activities was highly touted earlier in the decade, and those mountains of digital data are still tempting for federal spy agencies. But the council's 352-page October 2008 report, titled Protecting Individual Privacy in the Struggle Against Terrorists, concludes that such efforts will never amount to an effective way to get the bad guys.
The report was produced by a committee that included Charles Vest, the former president of MIT, and R. Gil Kerlikowske, Seattle's police chief. The authors' ire is aimed not at data mining directed at specific suspected individuals, but rather at broader scanning of the general population aimed at finding scary patterns. "If it were possible to automatically find the digital tracks of terrorists and automatically monitor only the communications of terrorists, public policy choices in this domain would be much simpler," they write. "But it is not possible to do so."
The findings suggest that Total Information Awareness-an effort launched in early 2002 to perform broad, warrantless data capture and analysis-was misguided from the start. The program generated a hue and cry when the public became aware of it, and Congress officially defunded it in 2003, but many of its core elements have continued in different agencies under different names.