Search Engines of the Future

One of the most exciting developments in search engine technology is the use of user preferences to narrow the search field. In part, this technology already exists: A personalized Google toolbar, for example, is a data collection device, and the “clustering” search engines such as Mooter refine a user's search by gradually honing in on the user's desires. The most useful image of this kind of search, however, may be the way the Amazon database functions. Amazon uses artificial intelligence to greet you by name. The database remembers past purchases, as well as what items you have viewed recently, and it offers customized recommendations based on past choices. Finally, Amazon produces a customized home page listing your favorite categories and offers you related media suggestions such as CDs by an artist to whom you have listened in the past or the latest book by your favorite author. Amazon also recommends products that have been bought by people who have bought the same things you have. Additionally, it posts reviews of products, both from experts and from other customers, and invites responses to both. Amazon contains your entire search history and uses that history—either to help you or attempt to sell you things—depending on your viewpoint. Years ago, Amazon was only a database, but it is now linked to the A9 search engine, which uses clustering and remembers your search history.

Google's motto is “organizing the world's information and making it accessible.” To that end, the company has gradually added a variety of specific functions to its search engine applications, including Froogle, a local shopping search engine; Google Maps, which provides maps and directions; Google News, a news service that users can customize; and g-mail, Google's e-mail service.

What else does the future of search engines hold? Imagine “searching” for your lost luggage from your PDA or cell phone, which a search engine will be able to find because your suitcase will be tagged with a microchip. Imagine a world in which everything is wired and everything is linked. We will go from using our phone to look up television sitcoms to being able to reset the air conditioner in our homes before arriving home, and checking in with our parents or siblings on our video wristwatch.

Potentially everything in your clickstream—every Web site you've ever clicked on—and the music you've downloaded, the videos and television programs that you've watched, every e-mail you've ever sent or received, and the hard drive of your computer and everything stored on it could be combined into a single database. Put it together and you get, potentially, either a search engine that would be very helpful in guessing what you want or a serious invasion of your privacy. It depends on who has access to the information and what uses will be applied to the data.

It's not as if someone is reading our e-mail and Web searches in real time, but it is being crawled for keywords that advertisers have bought. When Google launched g-mail, for instance, having failed to warn people that small advertisements would appear next to their mail based on its content, users were extremely displeased.

From a research point of view, this sort of keyword identification system merely raises questions about what happens when you become interested in new things. In some cases, users may be able to request that the computer simply ignore past preferences. From the standpoint of privacy, it results in legal issues, which will no doubt employ lawyers, if not computer engineers, and the need for more software that allows people to browse the Internet anonymously, using, for instance, anonymizer.com. It might also make us hope that the Web sites that collect information about us are very, very secure.