Last Updated on October 14, 2020 by Sean B
In this article, we’ll take a look at who invented the chatbot, but before we do that, I want to provide a little background. Chatbots are a fascinating and genius invention of the 21st
century and have reshaped our AI domain and our very intuitions about what machines can do. Chatbots are a prime example of smart AI machines that can successfully imitate and mimic human behavior.
Chatbots are used in wide arrays of spheres today, such as in entertainment, health sectors, finance, customer satisfaction, products and services, sales, educational platforms, and social media.
The AI Chatbots provide a very interactive and authentic way of communicating with human agents whilst mimicking human conversations. Chatbots helps human employees to focus on other issues whilst performing more mundane tasks. So, who invented Chatbots? Let’s look at the brief history of Chatbots and their inventors.
A Brief History of Chatbots
Suppose we want to trace back the history of Chatbots and identify the possible antecedents towards Artificial Intelligence (AI) development. In that case, we have to start off with none other than the pioneer in the field of computer science and mathematics, Alan Turing.
Turing’s Contribution Toward the Development of Chatbots and AI
Alan Turing (1912-1954) was a British mathematician, logician, cryptanalytic, and computer scientist. He is considered one of the most remarkable scientists of the 21st century and possibly the greatest mathematician of Britain.
Alan Turing laid the foundations of not only computer science but also AI. Alan Turing accomplished a number of feats, perhaps the most famous is breaking the Enigma code in World War II which saved millions of lives.
But let’s get back on topic, which is whether or not Alan Turing can be called the Father of Artificial Intelligence and therefore chatbots.
Universal Turing Machine
It has always been a dream of scientists to create machines that can mimic human behavior. Alan Turing authored one of the most important early works in the domain of AI. In 1935 Turing described a hypothetical machine capable of housing a limitless memory and a scanner that can move back and forth through the memory, symbol by symbol, reading what it encounters, and as well as write new symbols.
This scanner’s tasks would be dictated by an instructive program that is stored in the memory of the machine in the form of symbols. This hypothetical concept also implicitly included the machine’s possibility of improving and/or modifying the program by itself.
This abstract concept of Alan Turing came to be known as the Universal Turing machine, and all modern-day computers are Universal Turing Machines. Even during the war, Alan Turing devoted some of his efforts towards developing machines that could be deemed intelligent.
Conception of Problem Solving Machine
As a Turing colleague at the Bletchley Park, Donald Michie recalls in his anecdotes that Turing often discussed with him how computers could learn from experience and solve problems; now, this process is known as heuristic problem-solving.
In 1947, Turing gave a lecture on Artificial intelligence in which he described how machines could one day learn from experience and how letting computers modify their instructions could make all of this possible.
Turing proposed methods such as chess to determine machines and computers’ intelligence as chess is a challenging game that requires a lot of problem-solving abilities and intelligence. A chess-playing computer will look for the best moves to win the game by searching among the possible moves.
This was impossible in principle, as this required the machine to examine a large proportion of numbers. This method necessitated an authentic use of heuristics to guide the machine, and the conception of chess-playing machines remained theoretical for over 50 years.
However, in 1997, Deep blue, a chess computer that was built by IBM, defeated the reigning world chess champion Garry Kasparov, in a six-game match. Thus Turing’s prediction came true. Anyhow, this extraordinary feat is attributed to advancement in computer Engineering rather than AI.
The Turing Test
In 1950, Turing rephrased much of his intuitions and questions about Artificial Intelligence and introduced a Turing test to determine Artificial Intelligence machinery’s capabilities. Turing’s conception of his test can be summed up in his following quote:
“A computer would deserve to be called intelligent if it could deceive a human into believing that it was human.”
The Turing Test consists of three components, an intelligent machine, a human interrogator, and a human agent. The human interrogator’s job is to hold a conversation simultaneously with both the participants and find out which one is the human and which one is the machine.
The conversation requires the interrogator to ask a plethora of questions to both the participants, and in response to it, the machine should answer with statements that can trick the interrogator into thinking that they are talking to a human agent.
By application of the aforementioned principle, American Philanthropist Hugh Loebner introduced an annual competition that promised a $100,000 prize for computers that can pass the Turing test and a $2,000 prize for computers with the best effort.
However, to this day, no computer has ever truly passed the Turing test.
So, Alan Turing helped develop the concepts of modern computers, artificial intelligence, and the chatbot. But can he be credited with inventing chatbots? I’m not completely sure.
Then Who Invented the Chatbot
As for who actually invented the chatbot, that’s a difficult question to answer. Chatbots are a combination of technologies, so the inventors of each of those technologies can be credited with adding to the development of the chatbot. The early milestones in the field of AI were achieved before there were chatbots, and there were successful AI programs as earlier as 1951. And of course, Turing’s own contribution to the development of chatbots is huge.
Early Examples of Chatbots
Each of the following chatbots added something to their development. I believe that the first entry below is by far the most important however.
Eliza is known as the first chatbot ever developed in the history of computer science. Joseph Weizenbaum developed it at Massachusetts Institute of Technology (MIT) in 1966. Eliza employs the principles of pattern matching and substitution methodology to stimulate conversations with humans.
Eliza was great at mimicking conversations and worked bypassing the words users entered into the computer and then paring them to a list of already stored scripted messages, i.e., giving the most relevant scripted response to the human.
The script Eliza used stimulated a psychotherapist and proved to an enormous on the understanding of how language is processed. Eliza was great at providing the illusion that the human is interacting with another human and not a machine.
Learn more about ELIZA:
Chatbot History: The ELIZA Chatbot
Another capable chatbot was created by American Psychiatrist, Kenneth Colby, in 1972. The program mimicked a patient suffering from schizophrenia. It stimulated the condition of the diseases and aimed to mimic the behavior of an individual with schizophrenia.
Parry worked by a complex mechanism that was based on a system of assumptions, emotional responses, and attributions, which can be triggered by manipulating the weights assigned to verbal inputs.
The effectiveness of the Parry bot was tested by employing a variation of the Turing test. The bot was allowed to interact with human interrogators via a remote keyboard. These early tests were compared to the test models we have now.
Kenneth Mark Colby’s research was based on how computers can help understand mental illnesses; he also began a project “Overcoming Depression,” which carried on until his demise in 2001.
Learn more about PARRY:
Chatbot History: The Parry Chatbot
Next in the list is Jabberwacky, which was really a family of chatbots developed by British developer Rollo Carpenter in 1988. The developer aimed to create a chatbot that can stimulate natural human conversations in an interactive, entertaining, and humorous way.
Rollo carpenter aimed to design this bot for learning from human interaction. The developer intended to create this bot so it can eventually pass the Turing test, and for this purpose, the bot also participated in the annual Loebner competition.
Jabberwacky has no other functions other than mimicking human conversations since it’s conception; however, some individuals have sought its uses in academic research purposes. The bot was intended to provide human users with a humorous and natural conversation partner.
The developer aimed to see his bot eventually being used in home machinery and devices such as talking pet bots or other devices that can provide companionship to human users. Jabberwacky evolved in his own way through several Avatars, the most famous being George and Joan. Sadly, these earlier Avatars are no longer online, but you can try to interact with them through the Internet Archive.
Jabberwacky has since evolved into Cleverbot, who has several different Avatars of his own, including Evie the Cleverbot.
Learn more about the Jabberwacky Chatbot and Family:
Chatbot History: The Jabberwacky Chatbot and Family
Alice was developed in 1995 by Richard Wallace. Alice Chatbot is capable of natural language processing, which means that it does not need words to pair with pre-scripted messages in the memory.
This quality of Alice allowed for more interactive, authentic. And, sophisticated conversations between the human agent and Alice! Alice works with XML schema, which is also known as artificial intelligence markup language (AIML), which acts as a dictator to instruct the bot to hold conversations.
Alice is capable of performing many feats, such as telling the user name, age, hobbies, as well as provide information on topics. Today Alice is available in open-source, and users are free to create their Chatbots by using AIML powered by Alice.
Learn more about ALICE:
Chatbot History: The ALICE Chatbot
Mitsuku is a record five-time Loebner prize winner and was created by Steve Worswick using AIML technology. The bot looks and acts like an 18-year-old female from Leeds, England. She is regarded as the best Chatbot in the world and is capable of responding to questions that require critical thinking, such as can elephants fly.
The AI platform Pandorabot powers the bot, and the platform offers users to create their Chatbot using AIML technology.
Learn more about Mitsuku:
Chatbot History: The Mitsuku Chatbot
Well, who invented the chatbot? While Alan Turing never actually built a working chatbot, or any other Artificially Intelligent agent, I would argue that without his contributions, computers and everything that has developed from them would not exist. So at a minimum, Turing should at least be recognized as the granfather of modern chatbots.
Since we’re removing him as being completely responsible, I think we need to acknowledge that Joseph Weizenbaum deserves credit for creating ELIZA, the first true chatbot. While she was limited, mostly due to the technology available when she was created, everything that came after was standing on her digital shoulders.
So who created the chatbot? It is my belief that both Alan Turing and Joseph Weizenbaum should be credited with the creation of chatbots. Without Turing, the concept may never have been thought of, but without ELIZA, the first chatbot, no leaps forward would have been made.
I hope that the blog was helpful and informative.
I look forward to reading your thoughts about the article in the comments below.