#Eliza chatbot code#

The code can easily be modified for more advanced logic beyond the current a AND b cases. For example, “Apollo” AND “11” could give “Columbia/Eagle was launched with crew Aldrin, Armstrong, Collins”.

Logic: Generating a response if a pre-determined logic combination is found in the input.

Dictionary: Combining a starter phrase with definitions so an input with the adjective “arduous“ gives “In what ways is it difficult and tiring?”.

Keywords: Finding a keyword such as “joke” in the input produces a response such as “Why was 6 afraid of 7? Because 7, 8, 9”.

Simple Parsing: Changing 'My head hurts' to 'Why do you say your head hurts?“ by adding a starting phrase and switching pronouns.

Generic : Using a canned response such as “Can you elaborate on that?”.

The picture below shows the program flow for the following types of responses which are generated: It works by first splitting inputs into discrete words which are processed to generate responses. The Scratch program “Eliza” sprite contains the natural language processing and response generation code. We extended the basic ideas of the original Eliza by using a dictionary of thousands of keywords, nouns, verbs, and adjectives to expand the richness of the responses and provide knowledge on specific topics.

In the case of Eliza, its behavior is pretty straightforward to follow and you can quickly start to think of ways to make it better. Clarke once said that any sufficiently advanced technology is indistinguishable from magic. You can try it out and easily make edits at:Īrthur C. A screenshot of the Scratch project (which runs best in a computer browser) is below. To interact, simply type in responses and it will have a conversation with you. It can be easily extended to include knowledge about other topics and more advanced interactions. Similar to the classic Eliza, it acts as a Rogerian therapist but can also answer questions about Minecraft, Fortnite, and human spaceflight by leveraging a database of more than 8000 keywords and associated responses. Eliza Bickford is a modern implementation of the original chatbot which uses Natural Language Processing and artificial intelligence to have a conversation with its user. I decided it would be fun to make a modern version of this with my 10 year old son as a project for the two of us to do together. It probably didn't hurt that a new mysterious girl would seem intriguing to almost any high school boy. It was a bit like they were talking to a crazy person they couldn't quite make sense of.

The program was not always accurate, but I found that the user would simply ignore these and move on. The following year I was tricking friends into believing they were talking to a real girl over a modem when they were really just talking to the computer. I was first exposed to Eliza in the summer of 1992 where I created a simple version of Eliza in Scheme for an introductory computer class. that extremely short exposures to a relatively simple computer program could induce powerful delusional thinking in quite normal people.' Users unconsciously assumed Eliza's questions implied interest and emotional involvement, even when they knew that the program did not have real emotions. The basic program script was able to elicit emotional responses from users who began to perceive understanding and motivation in the program's output. Despite its simplicity, Eliza was remarkably effective. The ground work extends even further back to the 1950s, but computational limits kept a real chatbot far from anything most people, or my 10 year old self, would get to experience even in the 1980s.Įliza simulated conversation by using pattern matching and substitution to give users an illusion of understanding on the part of the program. Using natural language processing to interact with a user can all be traced back to the Eliza program which was created between 19 at the MIT Artificial Intelligence Laboratory by Joseph Weizenbaum.

Eliza chatbot