When AI research started in the 1950s, all the pioneers took human-level AI as the goal. As most of us realized, understanding intelligence well enough is a difficult scientific problem. Too much AI these days is oriented to short term applications, so it is time to spend some time thinking about the hard problems that lie between where we are today and the goal of human-level AI. We can divide the problems into epistemological and heuristic. Epistemological problems concern what knowledge is available about the world and how it can be expressed and what are the correct ways of reasoning about it. Facts about space, the consequences of action and knowledge present some of the epistemological problems. The epistemological problems come into focus when we consider the common sense informatic situation. An intelligent system needs to use approximate theories involving objects that are only approximately defined. When new phenomena become relevant facts about them need to be discovered and added to those being taken into account. The heuristic problems arise as soon as knowledge is adequate to determine a path to goal but it is necessary to find a path through the combinatorial maze of possibilities. The lecture will discuss a number of problems of both kinds and suggest directions for further research. The emphasis is on logical AI, but the problems arise also in biological approaches to AI. These problems are most likely to be precisely formulated and solved by people just now starting their work in AI.
John McCarthy is Professor of Computer Science at Stanford University. He has been interested in artificial intelligence since 1948 and coined the term in 1955. His main artificial intelligence research area has been the formalization of common sense knowledge. He invented the LISP programming language in 1958, developed the concept of time-sharing in the late fifties and early sixties, and has worked on proving that computer programs meet their specifications since the early sixties. He invented the circumscription method of non-monotonic reasoning in 1978. His current main research area is formalizing common sense knowledge and reasoning. His articles are on John McCarthy's main web page. McCarthy received the A. M. Turing award of the Association for Computing Machinery in 1971 and was elected President of the American Association for Artificial Intelligence for 1983-84 and is a Fellow of that organization. He received the first Research Excellence Award of the International Joint Conference on Artificial Intelligence in 1985, the Kyoto Prize of the Inamori Foundation in November 1988, and the National Medal of Science in 1990. He is a member of the American Academy of Arts and Sciences, the National Academy of Engineering and the National Academy of Sciences. He has received honorary degreees from Linkoping University in Sweden, the Polytechnic University of Madrid, Colby College, Trinity College, Dublin and Concordia University in Montreal, Canada. He has been declared a Distinguished Alumnus by the California Institute of Technology.
Time and Location: return to the 2000 Colloquia Series main page.