Artificial Intelligence: An Introduction

The science of making machines do things that would require intelligence if done by [people]
(Minsky, 1968)
For those of you who are new to Artificial Intelligence there are many introductions to be found on the Internet. I recommend this one. If you want to know more generally about AI resources on the Internet then start here.

What is Artificial Intelligence all about ?

Broadly speaking Artificial Intelligence is the study of the development of computer systems which emulate the various aspects of human cognition. These means developing computer hardware and software so that computers can reason, learn, see and understand what they see, hear and understand what they hear, speak, remember and recall.

There is some confusion about what constitutes Artificial Intelligence. This confusion arises because on the one hand there is an artefact called an Artificial Intelligence and on the other hand there is the subject, or area of knowledge, called Artificial Intelligence.

Artificial Intelligence: the artefact

We come across AIs quite a lot these days in science fiction films and books: HAL the computer in 2001, the robots in "I, Robot" written by Isaac Asimov and the AIs created by William Gibson in his books such as "Neuromancer". Whether you think these really are AIs is up to you to decide. Most people will have some idea of what constitutes an AI, things like: - ability to learn - able to have a reasonable conversion on a variety of topics You may require an AI to be self aware or conscious. Unfortunately, there are no objective tests for these concepts yet, so building an AI which possesses them is rather difficult. How would you know whether you had succeeded ? Others of you may contend that it is not possible to build an AI because there is some aspect of intelligence which it would need to embody, such as consciousness or have a soul, but which it cannot embody because it is simply not possible for a machine to possess this quality. All of these issues make it difficult for us to agree on what what actually constitutes an AI. Fortunately, for workers in AI, these issues are not necessarily important, though they are interesting to speculate upon. Most workers in AI are primarily interested in doing research and development on AI tools and techniques.

Artificial Intelligence: the subject or field of knowledge

The field of Artificial Intelligence has been around for about 50 years now. What comprises the field is best described by saying that it is what is taught in university AI courses, what is written about in AI books and journals, and what is presented at AI conferences. A typical textbook purporting to be an introduction to AI would include topics such as:

Expert systems, Deduction and theorem proving, Knowledge representation, Program languages, Learning, Natural language processing, Problem solving and search, Robotics, Vision and scene understanding.

These topics are not obviously related to the emulation of human cognition, but you will find that the successful emulation of many cognitive skills is often facilitated by some fairly esoteric (frequently mathematical) underpinning.

Last updated 30 July, 2002