This article that I found on New Scientist talks about intelligence gathered by AI software called KARNAC (Knowledge- Aided Retrieval iN Activity Context). KARNAC was created by Applied Systems Intelligence, Inc. KARNAC can sift through data located in a Global Information Base (GIB) to analyze for patterns of suspicious activity and match on seemingly unrelated events to predict future terrorist activity. The article says that if such a system had been in place that possibly both the Oklahoma bombing and the World Trade Center bombing could have been predicted. Of course, there are privacy concerns (privacy a casualty?), but it would be cool if the system could stop terrorists. Here's a press release from ASI. This of course is different than the Software intelligence agents reported about before on robots.net.