Scenario 1: a trustworthy local businessman reports suspicious activity by an apparently Middle Eastern neighbor.

A citizen contacts a detective in a small east coast city. The detective knows the caller to be a trustworthy local businessman. He reports suspicious activity by a neighbor who moved into a rental residence nearby. The neighbor appears Middle Eastern. The neighbor claims to be a student at the local university; however, he is noticed to be absent from the residence for weeks at a time. The caller engaged the neighbor in conversation and learned that although he claimed to be enrolled in an International Studies program at the University, he was very vague and clearly did not recognize the names of the most prominent professors within that program. The neighbor has observed three other Middle Eastern males intermittently staying at the neighbor's home, sometimes when his neighbor is not there.

Problem: This scenario was viewed as quite typical of the many that have come through local police and FBI offices since 9/11. The problem is how to develop an effective triage system that helps officers or agents handle the large volume of incoming information while assuring that important details are not overlooked.

Strategies: Make use of data gathering/vetting systems already in use in other situations, such as in the medical and legal professions. These are designed to (1) process all incoming information and then sound an alert when a targeted item appears, and (2) show changes in patterns of data flow that would not be seen by the casual, part-time observer (an individual on duty at any particular time would see only that part of the incoming data, whereas the artificial system sees it all). The pattern of change in incoming data might be informative, in addition to targeted items.

For example, one such system already in use searches through large volumes of text for specific words or phrases. When it finds a targeted item, an alarm is sounded. An important aspect of this system is that it will find text that means the same without using the same words. For example, if a target search specifies "racketeering," one will get documents that mention racketeering, but also documents that mention "unlawful conspiracies." (1) Also, the data are kept in files that can be re-searched for words or meanings that are of interest at some later date, to look for similar instances, as well as to allow for the generation of graphs and other descriptors to evaluate changes in the pattern of incoming information.

Further information is provided in Appendix 1 (Information Evaluation Systems).

Problem: What predictor variables do we have for the identification of potential terrorists?

Strategies: We might consider using the information that is known about the men who attacked New York and Washington on 9/11. For example, it appears that the men who attacked on 9/11 had not maintained their family relationships, perhaps in order to ensure good cover. Thus, the presence or absence of family relationships might be used as predictors of membership in terrorist networks. Such information also might be useful in the interrogation and identification of suspects.