G5AIAI - Introduction to Artificial Intelligence

This course is run at the The University of Nottingham within the School of Computer Science & IT. The course is run by Graham Kendall (EMAIL : gxk@cs.nott.ac.uk)


Eliza

Eliza (Weizenbaum, 1966) was one of the first AI programs and, at the time, it impressed the people who used it.
The name Eliza was chosen as "it" could learn to talk increasingly well as did Eliza of Pygmalion fame.
You asked Eliza questions and she (it?) responded.

Here is a sample script from an Eliza session

Person : Men are all alike.
ELIZA : In what way?
Person : They're always bugging us about something or other.
ELIZA : Can you think of a specific example?
Person : Well, my boyfriend made me come here.
ELIZA : Your boyfriend made you come here?
Person : He says I'm depressed much of the time.
ELIZA : I am sorry to hear that you are depressed.
Person : It's true. I am unhappy.
ELIZA : Do you think coming here will help you not to be unhappy?
Person : I need some help, that much seems certain.
ELIZA : What would it mean to you if you got some help?
Person : Perhaps I could learn to get along with my mother.
ELIZA : Tell me more about your family

Eliza was based on a set of rules that it used to match the left hand side of the users input. Here are some examples of the rules that ELIZA used

(X me Y) ---> (X you Y)
(I remember X) ---> (Why do you remember X just now?)
(My {family-member} is Y) ---> (Who else in your family is Y?)
(X {family-member} Y) ---> (Tell me more about your family)

By looking at these rules you can see how some of ELIZA's responses in the sample script were generated. You should also be able write down some more rules if you have the inclination.

If ELIZA is able to match two or more rules then its response will be chosen at random. This adds to the illusion that ELIZA has some understanding.
In addition, some of the rules have no left hand side. In effect, they will match with anything that the user types in. These rules can be used when no other rules match. This can lead to a response such as

Can you elaborate on that for me?

Or

That is very interesting. Why do you say that?

Eliza was designed to work with different "scripts." That is, the program could talk about different topics depending on the rules (script) it was given. Weizenbaum decided to use a psyhotherapist in the first instance as many responses are simply re-phrasing (or turning around) what the patient has said. Therefore, it was quite easy to implement - rather than trying to get Eliza to hold a discussion on (say) politics.

There have been reports of people becoming so dependent on Eliza that the author eventually had to withdraw its use.
Another story tells of the author's secretary using Eliza when he entered the room. The secretary asked her boss to leave until she had finished her conversation.
The author also considered putting in a module that recorded peoples conversations. This was greeted with accusations that Weizenbaum was spying on their secrets and innermost thoughts.

In fact, such an impact did the program have that some psychiatrists saw it as a means of allowing the profession to deal with many more patients. There was talk about using Eliza as a front end with the most serious patients would be referred to a human psychiatrist. Despite Weizenbaum's insistence that the program did not understand the patient's problems, some in the profession still did not believe him.

And, all this about a program that is simply text matching and cannot be considered as having any understanding.

If you want to try an "ELIZA session" there are several programs around and a couple of web based versions are given below.

You might also like to read (Weizenbaum 1993). This book is written by the author of Eliza. The book is really an explanation of how computers work and what they can and cannot do but he does discuss Eliza as well as AI topics.

 

Place to Visit

References

 


 Last Updated : 11 Sep 2001