ven if you are a limited artificial intelligence agent, you have probably noticed that this issue of the Bulletin is devoted to the special topic of Agents. Anyone knowledgeable about the history of AI cannot help but recognize that yet another neat little technology is being hyped nearly to death, and anyone knowledgeable about "agents" can not help but recognize that, as an artificial intelligence technology, there is typically very little actual intelligence involved. In some sense, Agents can be defined as small scale AI that works. This was particularly evident at the First International Conference on Autonomous Agents (Agents-97) in Marina Del Rey this past February. There is a lot of activity and real research going on related to Agents that is fascinating and occasionally entertaining. Agents-97 attendees certainly noticed the most popular presentations were those given by or relating to the entertainment industry. This seems perfectly natural when you think about it. Who better to capitalize on this new wave than Hollywood? First, they are the only industry that is better than AI at creating hype: on this front we seem natural allies (or adversaries). Secondly, is there anything on Earth that encapsulates less intelligence than the average American TV sitcom? Agents looks like encyclopedias in comparison! What a perfect union. The overlap between Agents and Entertainment does not stop there, however. Entertainment and AI have always had great visions. Think of the power displayed in The Day the Earth Stood StilL" Gork provided the best demo ever to be seen inside the Washington DC Beltway. Remember Sneakers? What about Goldeneye? These movies clearly cry out for a huge new area of research and development E that has thus far been overlooked: Autonomous Secret Agents (ASA). These are to be followed by a second generation of intelligent autonomous secret agents, replacing the existing human-based central intelligence agents. At Agents-97 there were numerous calls for various new areas of Agent research. The keynote speaker called for a "language of emotion" while others called for x, y and z. Clearly the omission of a call for research on ASA will go down as one of the great scientific blunders unless we act immedi- Board Information Warfare - Defend, http://jya.com/iwd.htm) along with all the news from CERT (The Computer Emergency Response Team, http://www.cert.org). One of the obstacles to the development of the autonomous secret agents field is the lack of any real foes. The, demise of the cold war has left the west without a readily identifiable evil enemy, and so far government sponsored research has not met the level of the potential threat now verified by the consistent recommendations of the Defense Science Board on IA/IW-D. th the real 3w full well :i to do with luch of the SPECTRE, le infamous dng havoc r James Bot search must rely, before destroy the i c e , not to while conyour local boracion of nse indus[itary-enter~d to develkSAs. ire enjoying ~f SIGART configuring our page at ~u have any ; column or the mag,=elfree to send them to backtracking@sigart,acm.org. • (logs), reactive (at least hours after anything is noticed) and manual (operation notification and intervention). If we divulged any more information, we'd have to cancel your subscription. Have your local web agent check out such sites as IDA (The Institute for Defense Analysis, http://www.ida.org) and DSB IW-D (The Defense Science Chris, Wel'r,y and LOU Hoebel 199'7.';51GGAPRT Bulle'r..in ii

/lp/association-for-computing-machinery/backtracking-EGz7bBR089