Agents of Today and Tommorow

Cool stuff, or all just fluff?

Agents Today

Search Agents

There are currently many different utilities which will allow you to search through thousands of URLs to find specific information. A list of the more popular ones can be found here.

Of these searching utilities, several of them fall under the category of WebBots. WebBots are programs which:

* Have a query-search utility for a database with thousands of URLs
* Have a database which is generated by an agent which wanders the Web in search of new URLs

Two of the more popular WebBots, WebCrawler, and US government database has started to take drastic measures to prevent these attacks. Just please don't click on the "seek and destroy" button.

Ideas have been put forth as to ways of preventing these attacks, but no one standard has been agreed upon as of yet, although one proposed standard has wide-spread support in many WebBots today.

Personalized Agents

Two personalized Agents which can be used today are being developed at the MIT Media Lab. Each of these agents tries to learn about you, the user, by asking questions. Your answers to the questions are then used, and compared to other users, to try and learn your preferences.

The first of these agents is WebHound. Like HOMR, WebHound tries to figure out your tastes, only this time it is trying to figure out what kind of Web pages you like to read. You can give WebHound your hotlist, and then rank each item in it on a scale from 1 to 7. WebHound will then try and find some URLs that you might be interested in reading. Just imagine the time you can save in Web Surfing alone! Again, you need to subscribe to the service, but it is worth it. You can also put a bit more effort in and set up WebHound as a client-server architecture. You can get daily recomendations on new Web pages.

If you would like to try out some more Agents, or just want to learn a bit about agent architectures, surf on over to the Intelligent Agents List. This is a list of many other agents, from theoretical models to working products.

Agents in 5 years

It seems that every time a prediction is made as the where computing technologies will be in 5 years, they are undoubtably proven wrong when the five years are up. But I'm going to make a few predictions as to the progress of Intelligent Agents anyhow.

Apple's "Intelligent Agent for 1992", Phil, has obviously not been developed yet, and it is three years after that date. However, even with current technologies, this "almost-intelligent" agent can be somewhat implemented. Improvements in voice-recognition technologies are the most obvious weak point in this type of system.

Even without 100% voice-recognition in the next five years (which seems unlikely), we may see these primative agents appearing on desktop computers, eliminating the need for users to perform repetative, mundane task over and over again. Perhaps a 'desktop agent' will allow us to finally find that progress report from last year which has disapeared somewhere onto that hard-drive, never to be seen again...

Agents in 10 years

It is quite likely that in 10 years, voice-recognition technology will have progressed to a point where it is used in every-day computing. With such an advance in technology, we may begin to see more Intelligent Agents available for use.

Apple's 'Knowledge Navigator' vision may seem far-fetched at first, but may actually be partially developed in this time frame. The ability for a computer to interact with a student to present information and teach new topics would have a dramatic effect on the educational system. An Agent could be much like a personalized tutor, noticing when the student is having troubles with a topic, or when they might be able to skip ahead to newer material. Personalized education could well be on it's way into our lives.

Agents in 20 years

It is impossible to tell where Intelligent Agent technology will be in 20 years. Perhaps we will see devices like AT&T has promised to deliver, which will put a personal assistant inside of our notebook computer.

Or, perhaps, we will not have Intelligent Agents in the future at all. Although there are several companies which are pushing Agent visions, (AT&T and Apple are the most noteworthy), it may turn out that Agents just don't become popular. Or maybe there is an unseen technology barrier that we will come upon in the next few years that will prevent such agents form being developed.

Whatever the outcome in 20 years, there is no way to predict today what it will be like then. All we can do is guess...

* Wander on back to Angela's Building Interface Agents page.
* Head on back to Mark's Introduction page.

Back to report list