Intelligent Agents: It's Nice to Get Stuff Done for You

Article excerpt

'Intelligent agents"-chances are you've heard plenty of tall tales about these super-cyber entities that are fast appearing on the Web scene. You'll hear them talked about using a variety of different nouns and phrases. People describe them using such terms as smart agents, search agents, Web robots, bots, avatars, intelligent assistants, smart applications, customer service robots, shopbots, intelligent tools, and other names.

Bots are not new ideas; they are instead predictable developments based on the Artificial Intelligence technology that flourished in the 1980s and 1990s. One of the first articles to recognize the importance of this technology to information professionals was written bv Marina Roesler and Donald T. Hawkins in 1994 ("Intelligent Agents: Software Servants for an Electronic Information World (and More!):

Intelligent Agents Help Their Users with Routine Computer Tasks, While Still Accommodating Individual Habits," ONLINE, July 1994, pp. 18-32).

What are these tools, and what can and can't they do? Is this is just more hype, hogwash, and sales pitch overdrive? Admittedly, it does sound like science fiction, but on the other hand, these tools really can be the 21st century magic solution to information retrieval problems. Or, is all this in the same fantasy vein as the supposed new ability to cut down on library and research staffing and funding, "since it's all out there on the Web for free"?

JARGON IN, CONFUSION OUT

Part of the reason for the confusion is the baffling use of jargon to describe these software tools. Yet these are all variations of the same "smart software" technology vision of the Web. The point of Web smart agents is to turn the network into a more interactive place, with easy location and transferal of information between native Web applications and free-ranging software agents. This information is then ready to be further interpreted and transformed by local processing. The end result is that the Web will also be a friendlier and more inviting environment.

Such intelligence leads us toward what some visionaries are calling the Semantic Web. This is a network future where software agents roam from page to page-mining data, text, and ideas and performing sophisticated tasks at the bidding of their human masters. This vision includes ideas such as automatic "semi-intelligent" exchange of information between agent programs, intelligent agent software traveling across the network to actually operate on different machines, and use of speech communication and interaction between the user and the network. This isn't hype; you can already read about and use prototypes and demos of such tools. In the Semantic Web, agents will connect and harness the latent power of distributed network intelligence.

CONTROLLING EXTERNAL DEVICES

At this point, the next step is for agents to fully develop interconnections between the smart network and the external or physical control world. Then agents/programs can expand to advanced control abilities over external devices, such as electronic device microprocessors, operating controls, motors, or thermostats. Remember, we've already started playing with simple external control applications, such as remote Web operation of TV cameras, monitoring inventories of dormitory soft drink machines, using the X10 interface for control of "smart home" controls. This blurs boundaries between network smart agents and external world robots, in the Issac Asimov sense of the term.

You may remember the Apple Computer film of the late 1980s, depicting a human operator using a smart agent desktop/wall display to communicate and interact with the external world. That was a slick Star Trek vision of a sentient assistant to help in navigating the conceptual and physical universe and to perform complex tasks. Seems like this kind of Web may not be that far away.

The basic smart agent premise is practical and valuable. …