Grounded in reality: the infiltration of AI into practical educational systems

The paper aims to present a `bottom up´ approach to showing the relevance of artificial intelligence for systems to support learning. In the past, work in AIED has tended to be driven by researchers who begin with some idea or theory about some cognitive process (such as creative problem solving, qualitative reasoning, learning from analogies, and so on) or perhaps with some particular technique (such as assumption based truth maintenance, fuzzy logic, case based reasoning, etc.) which may have some use in educational systems. We then try to find a knowledge domain (such as geometry problem solving, medical diagnosis, foreign language learning, etc.) in which to explore these theories and techniques. And, finally, if we are lucky, we might try out our system in a realistic context. The author presents four case studies which are representative of the range of computer based learning systems. All were begun with a practical objective; all have evolved to discover the need for AI techniques. None is an `AIED system´: AI is not a major or motivating component of the system, it is just one of a battery of techniques or technologies which are brought to bear on the problem. Then the author gives some general comments on where AI fits into the design of practical learning and training systems