The Chatbot – Economic historical perspective about
Despite appearances, chatbot systems are not a current creation.
Economic historical perspective ideas appeared in the middle of the last century, when Alan Turing, one of the pioneers of computer science, wondered if machines could think and set the parameters by which a machine would pass as intelligence.
As computers began to develop, the first programs were able to simulate conversation based on voice or text commands in the 1960s. They have evolved with computing systems, but in recent years, thanks to advances in machine learning algorithms and neural networks, they have been able to receive a higher level of Artificial Intelligence, which has broken the ice of skepticism and opened the door to investment.
The first public notoriety project of a chatbot
was called Eliza, a natural language-based conversational program launched by Joseph Weizenbaum (Professor at the Massachusetts Institute of Technology) in January 1966. It ran on an IBM machine and sought to create the belief that the user was in dialogue with a psychotherapist.
Eliza was scheduled to ask open-ended questions and follow up on the data entered. The program recognized words or phrases in the text entered by the user and returned predefined answers. If the user said “I have a problem with my brother”, would Eliza continue the conversation with an open line “Tell me more about your family” or “My dog is dead” and continue “Your dog was important to you”? This created the illusion of a level of understanding and almost human interaction. The term chatbot did not exist in 1966, but Eliza aroused the imagination of a whole generation of researchers, writers, screenwriters, directors, etc.
Eliza was followed by other projects that attracted the public’s attention: Parry in 1972, Racter in 1983, ALICE in 1995, and Jabberwacky in 2005. The dialogue between Eliza and Parry, available here https://tools.ietf.org/html / rfc439) for example, remained famous, even though it reflected technological limitations and animosities among researchers. However, over time, the improvement of these systems and the achievement of a higher level of refinement of conversations are noticeable.
However, the switch to modern chatbots was made by IBM Watson, created in 2006 by the well-known IT company, which in 2011 managed to win the American Jeopardy contest against two of the best former champions of the show. For this success, Watson was trained for five years, living in 10 racks with Power 750 servers, which produced a high level of heat and noise, and the questions were addressed through a text interface. Watson later became the basis for the development of IBM’s Artificial Intelligence systems.
The ultimate goal in this field of activity is to build a system that supports a natural conversation with a human user. Recent advances in several areas of technology have brought us closer to this goal.
In recent years, chatbot
have gone beyond the poetic stage and entered the economy directly with the advent of virtual assistants. Siri, Google. Now, Alexa, and Cortana have shaped a new industry in which billions have been invested.
At the same time, chatbots have remained in the realm of entertainment, and the names Mitsuku, Red, Poncho, Right Click, Insomno Bot, Dr. AI, and Melody are probably the most famous in the online environment. Mitsuku, for example, has won the Loebner Prize four times (an annual competition in which chatbots are evaluated for the human nature of their conversations) and is considered the best conversational chatbot in the world.
An important moment for this market was in 2016 when Facebook launched a messenger platform through which developers can create chatbot interfaces for interaction with social network users. At the end of 2018, no less than 300,000 chatbots were active in Facebook messaging, three times more than in the previous year, and in 2019 the growth will be just as impressive.
The leap was made possible both by the huge footprint that FaceBook has in society (people exchange over 2 billion messages a month via Facebook Messenger) also by simplifying the process of developing chatbots. With over 6 million companies actively promoting themselves on Facebook, the opportunities are hard to imagine.