Where Should the Intelligence in Intelligent Interfaces be Placed?

Tom Kaczmarek, Robert Neches, John Seely Brown, Phil Hayes, Tom Malone, Dic Waters, Bob Wilensky, Mike Williams

Obviously, many AI techniques are applicable to building better human-machine interfaces. Intelligent interfaces have explored the use of problem solving, planning, heuristic search, discourse models, user models, knowledge representation, expert systems, and natural language text understanding and generation. The particular techniques researchers have chosen have led to two very different paradigms in intelligent interfaces. On the one hand, knowledge based, "intelligent apprentice" systems like the UNIX Consultant seek to provide assistance based on an understanding of the user’s intentions and task domain. On the other hand, "power tool" systems like COUSIN emphasize a powerful command set, but leave the responsibility for selecting and applying those commands in the hands of the user. Behind these distinctions are some very different fundamental assumptions about the problems that users need help with, and the AI techniques that can be applied to helping them.


This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.