The Community for Technology Leaders
Green Image
<p>In the future, intelligent software agents will help us navigate the information superhighway by serving as backseat drivers or taxi drivers. Better yet, they have the potential to act as sophisticated concierges who make it unnecessary for us to approach the highway at all.</p> <p><i>"The most profound technologies are those that disappear."-Mark Weiser</i></p> <p>Electric motors have exerted a dramatic influence on our lifestyle. There are, for example, more than twenty such machines in the average American home. But most people don't notice electric motors in daily life; instead, they focus on the activities the devices enable: washing clothes or listening to a CD. Similarly, computer technology has dramatically enhanced our ability to generate, deliver, and store information. Unfortunately, our tools for locating, filtering, and analyzing that information have not kept pace. Consequently, instead of "disappearing" like electric motors, the "information superhighway" looms larger and more prominent in the lives of its users every day. Intelligent software agents can reverse this trend.</p> <p>Such agents have two ways of making the information superhighway disappear:</p> <p><li><i>Abstraction:</i> Details of the technology underlying the agent and the resources the agent accesses are "abstracted" away-that is, they are user-transparent. The agent enables a person to state what information he or she requires; the agent determines where to find the information and how to retrieve it.</li> <li><i>Distraction:</i> The agent's character or persona helps distract the person from what might be a tedious or complex task.</li></p> <p><b>Forecast:</b> While cute agent personalities will have tremendous impact in the entertainment market, people will not appreciate them in operating systems, utilities, or business applications. Microsoft's "Bob" will become extinct for the same reason that consumers rejected talking cars.</p> <p>For intelligent software agents, we believe that abstraction will win out over distraction. The pervasive interaction style for today's computers is direct manipulation: the user clicks on, drags, and drops icons. Although direct manipulation is handy for performing simple tasks such as printing files or invoking applications on a personal computer, it does not scale to searching massive networks for information. Many visionaries recognize this point. Alan Kay writes:</p> <p>A retrieval "tool" won't do because no one wants to spend hours looking through hundreds of networks with trillions of potentially useful items. This is a job for intelligent background processes that can successfully clone their users' goals and then carry them out. Here we want to indirectly manage agents, not directly manipulate objects. (p. 203)</p> <p>Nicholas Negroponte uses the task of making one's bed as an illustration:</p> <p>Today, notwithstanding the skill, I cherish the opportunity of delegating the task and have little interest in the direct manipulation of my bedsheets(. Likewise, I feel no imperatives to manage my computer files, route my telecommunications, or filter the onslaught of mail messages, news, and the like. I am fully prepared to delegate these tasks to agents I trust as I tend to other matters...(p. 347)</p> <p><b>Facts:</b> The Information and Interactive Services Report states that more than 10,000 people are signing up for on-line information services each day (New York Times, April 9, 1995). The Internet Society estimates that more than 30 million people worldwide have access to the Internet on any given day (New York Times, May 11, 1995).</p> <p>Both Negroponte and Kay suggest that direct manipulation is appropriate for enjoyable tasks. Although some people enjoy browsing Internet content, most people would gladly delegate tedious tasks such as searching for information on the Internet to a competent agent.</p> <p><b>Fact:</b> There are at least 1,500 general-purpose information-finding services (staffed by human information brokers) currently doing business nationwide (Harvard Magazine, May-June, 1995, p. 23).</p>

D. S. Weld and O. Etzioni, "Intelligent Agents on the Internet: Fact, Fiction, and Forecast," in IEEE Intelligent Systems, vol. 10, no. , pp. 44-49, 1995.
189 ms
(Ver 3.3 (11022016))