Homeland insecurity: Datamining, terrorism detection

Following the terrorist attacks of September 11, 2001, there was heightened attention in the United States on the use of multiple government and private databases for the identification of possible perpetrators of future attacks, part of an unprecedented expansion of federal gov-ernment datamining activities, many involving databases containing personal information This article reviews some proposals that have surfaced for the search of multiple databases without compromising possible pledges of confidentiality to the individuals whose data are included and their link to the related literature on privacy-preserving datamining. In particular, we focus on the concept of selective revelation and its confidentiality implications. HOMELAND SECURITY AND THE SEARCH FOR TERRORISTS A recently issued report from the U.S. General Accounting Office [8] notes that at least 52 agencies are using or are using or planning to use data mining, “factual data analysis, ” or “predictive analytics, ” in some 199 different efforts. Of these, at least 29 projects involve ana-lyzing intelligence and detecting terrorist activities, or detecting criminal activities or patterns.1 Perhaps the most visible of these efforts has been the Total Information Awareness (TIA