Conversational bots are fashionable. All you have to do is look at companies like Google, Microsoft or Amazon, all of them investing in the development of a technology that allows you to simulate a conversation between people, automatically and coherently answering the questions you receive, either in a spoken way or written.
The software that gives life to bots, also called chatbots, relies mainly on two technologies: artificial intelligence and language processing. The combination of both makes it possible to recognize a series of words or expressions and respond using previously prepared sentences. In this way the bot is able to follow the conversation and answer naturally.
Being a current trend may seem that conversational applications are something new and yet it is not so, go back to the beginnings of computing itself.
In 1950 the British mathematician Alan Turing, a forerunner of modern computing, proposed a theoretical test that considered the ability of machines to show intelligent behavior, anticipating the idea that they would come to think. The test does not evaluate the ability of the machine to respond correctly, but to generate responses similar to those of a human.
Turing’s work inspired many computer scientists, among which the German Joseph Weizenbaum from MIT stands out. In 1966, he developed the ELIZA program with which he intended users to believe they were talking to a real person.
ELIZA: the first bot in history
Initially developed as a parody of the first interviews with certain psychotherapists, ELIZA was designed as a tool to show the superficiality of communication between people and machines. Thus, it used a sequence that recognized key words and generated an appropriate response, usually in the form of a question about how the subject in question made the user feel.
Professor Weizenbaum found five critical points to solve to ensure the proper functioning of his program: the identification of key words, the discovery of a minimal context, the choice of appropriate transformations, the generation of adequate responses and the ability to react in the absence of critical words.
The myth says that ELIZA, name taken from the street florist protagonist of the work Pygmalion (George Bernard Shaw, 1912), adapted to the theater as My Fair Lady, was able to pass the Turing test. This statement is somewhat risky and leaves room for debate, but what is certain is that many people thought they were talking to a therapist and shared intimate details of their lives.
This first bot was the inspiration for many others that developed later as ALICE, Mitsuku, Albert One or SmarterChild, one of the most popular.
From SmarterChild to virtual assistants
SmarterChild, developed in the year 2000, stood out for its ability to process the language, understanding and responding in a natural way to the questions that were asked. But it was also the first bot to provide help to the user, being able to provide information on a wide range of topics such as movie schedules, weather forecasts or sports results.
Compatible with instant messaging services such as AIM, MSN Messenger, AOL Instant Messenger and ICQ, brought artificial intelligence to millions of users, being in a way the forerunner of current virtual assistants, such as Apple’s Siri, Microsoft’s Cortana or Alexa’s Amazon
How far these bot-based services can evolve is still to be discovered. An example of this is, in the world of fiction, the experience of Theodore Twombly, a character played by Joaquin Phoenix, in the movie Her (Spike Jonze, 2013).