What was the first chatbot in history?

Written by Daniel Cerdas Mendez on Planeta Chatbot.

History of Chatbots and Virtual Assistants

You may like : How does a chatbot work?

  • ELIZA (1964): The first Bot, developed by Joseph Weizenbaum , at MIT for IBM 7094, considered the first bot, capable of english dialogue on any subject, Eliza, used tags to understand the texts and catalog them. In addition, it was set up to talk to users about their problems playing the role of a psychiatrist.

  • CHATTERBOT (1989): TinyMud, was a multi-user game about dungeons, which included multi-user conversations and simulated scenarios. In TinyMud, player controller computers were called Bots (as a short word for Robots) and were based on ELIZA. ChatterBot, is a TINYMUD virtual player that was created to chat with other players, explore the worlds, discover new routes through rooms, answer questions from other players about sailing. This Bot was successful in TinyMud, because the Turing test was applied, as all players assumed that all players were one person and did not know about TinyMud’s artificial intelligence.


ChatBots development for business. We create chatbots on WhatsApp, Facebook Messenger, Telegram… Contact us on our website.

  • ALICE (1995) (Artificial Linguistic Internet Computer Entity) It is a Bot inspired by Eliza, capable of collecting examples of natural language through the web. It used patterns to manage the conversation with the user, in addition, the data consisted of objects called AIML, which were organized into categories.
  • Clippy (1997) (Microsoft) The first conversation agent available on Windows. Microsoft designed it to help people used the Microsoft Office tool in 1997, however Clippy and his other personalities like the magician, cat, and dog disappeared in later versions of Office 2003
  • SIRI (2011) (Apple): It is the first virtual assistant for a phone created by Apple. Nuance is in charge of speech recognition, and together with Apple developers they managed to create a totally friendly conversation agent that responds to queries made such as weather, music, math calculations and history, and chip-chat conversations. Siri also uses the Google service as the engine to respond to queries from users who require web searches. Thanks to the Machine Learning behind Siri, its knowledge database increases according to the number of people using the virtual agent and uses the data provided as pronunciations, meanings and localities of language to improve the user experience.
  • Watson (2011) (IBM): It is an intelligent system created by IBM and is also recognized for competing in the Jeoparty program; it is a cognitive system that identifies language with the precision of a human being, faster and faster. Watson, is able to break down questions, create various answer hypotheses and respond with the highest probability.
  • Cortana (2014) (Microsoft): It is the Microsoft assistant that has expanded on cell phones, tablets, computers and video game consoles like Xbox all this through Windows 10, is based on an AI agent of a game called Halo, this assistant can learn and adapt based on Machine technology Learning, this leveraged by the Microsoft Bing Satori engine which is responsible for indexing millions of data.
  • Alexa (2014): This is the first virtual assistant created by Amazon, its main feature is that it can be used with the smart speaker called Amazon Echo and allows you to know information about weather, products, purchases, reminders, and even video calls.
  • Google Assistant (2016): The assistant created by Google, is available on mobile phones and smart speaker called Google Home, capable of including other Chatbots in your conversations, is a proactive assistant for addresses and business information.

Definition The word Chatbot, comes from “Chatterbot”, was a multi-user dungeon game, the main task was to answer user questions about navigating the dungeon and this agent simulated conversational skills by simple rules.

Read also : How to set up a Facebook chatbot?

Chatbots, are computer programs that can interact with users using natural language as defined by Bayan Abu. Wilks in 2006 mentions that they are an intelligent cognitive agent who knows the owner and his habits, speaks and assists with simple tasks.

Deen Allison mentions that Chatbots are conversational agents, artificial conversation entities capable of imitating human personality, interacting, and responding in statements to track a conversation meaningful to human beings.

Chayan mentions that Chatbots are programs that establish artificial conversions through text entries, they are used in customer service applications.

Therefore, Chatbots can be defined as cross-platform virtual agents capable of interpreting human language through voice, writing or capturing images and responding through a conversation, in addition, their interaction with people and other bots available. Chatbots can be easily scheduled and configured to perform tasks in the face of user queries.

Its operation consists of three ways:

1. The user makes a request via text or voice.

2. The request is analyzed by the agent and its artificial intelligence (BackEnd).

3. The agent responds in real time through a conversation.

In this image we can see the growth in terms of web search of the word Chatbot according to Google Trends over the period from 2012 to 2016.

Google Trends

Within the topics related to the search for the word ChatBot , Google Trends mentions: ChatBot Tay, Facebook Messenger Chatbot, Tay, Slack ChatBot, Chatbot Twich, TensorFlow, Kik, Alexa, ChatFuel, Deep Learning Chatbot, Racist ChatBot, Doulingo ChatBot and Xiaoice. This reflects the growth of development tools and bots available on the web.

There are also a number of concepts that should be understood when talking about Chatbot and artificial intelligence:

Machine Learning is a method of data analysis that automates the model to learn how to perform tasks or understand concepts and layers of generalizing behaviors thanks to the data provided. There are algorithms such as Decision Tree, Bayes Ships, Logistic Regression, SVM, Assembled Methods, and Cluster Algorithms.

Deep Learning allows computational models that are composed by multiple layers of processing to learn data representations, these learning methods have been improved thanks to speech recognition, object recognition and NPL.

Google Machine Learning

Cognitive computing , helps to develop a coherent, unified mechanism based on capabilities of the mind, whose goal is to mimic the functioning of the human brain using a computer model.

Application for Use

  • Medicine: For the China region, Microsoft released a Chatbot called Xiaoice used by around 40 million people and has recorded up to 10 billion conversations. It was designed by developers and psychological experts to create a balance between artificial intelligence and emotional intelligence. This Chatbot is able to memorize and track the emotional state of users and even offer a 33-day course of therapy for people with relationship problems. On the other hand, IBM uses its Watson service focused on medicine for different uses, such as: oncology, genetics, general medicine consultation and personal care.
  • Customer service: These conversation agents act as customer service representatives, giving answers in natural language and providing more focused information for conversation with a user . Chatbot is required to have the same tone, sensitivity, and behavior as a human service agent, but it is also required to process information faster than a human being.
  • Learning: This type of chatbots, based on conversations, facilitates the student online training in fields such as learning a second language, for example: TutorBot used the role of class assistant to provide services at any time and in addition, within his answers included the course reference materials, dictionaries. In addition, it facilitated conversations for the tutor to measure the progress of each student.

  • Entertainment: Skype has a series of bots, which allow you from consulting about movies, showing news, predicting the weather, to playing with the user chess and the popular game UNO. It also allows the creation of memes and trivia games.

There are also other categories, where Chatbots are venturing with new development platforms: data analysis, communication, design, development, education, file management, finance, food, health, technical support, human resources, marketing, news, payments, shopping, sports, travel and utilities.

Fashion Chatbot

Ethical considerations Certain amendments to the human rights statutes should be considered to include artificial intelligence and oblige them to respect those statutes, in order that robots respect each person or individual and in turn be respected in a human form.

Issac Asimov, in 1984 proposed three robotics laws that it should follow autonomous agents:

  • A robot cannot hurt a human and also a human cannot harm a robot.
  • A robot must obey the orders given by humans, except those that cause conflict with the former.
  • A robot must protect its own existence and its protection without a conflict between the first and second laws.

Therefore, it is the duty of developers and creators of virtual agents that their projects respect these three laws and include values such as: privacy, transparency, respect for human values, freedom and the common good.

Spam Chatbots

Different companies overused Chatbots in different chat rooms in early 2000, taking advantage of internet marketing and as a result spambot appeared such as: Hackers Bot, Scrapers, Spammer and Impersonators. Previously, a Bot operator controlled hundreds of Chatbots that sent links to thousands of users in different chat rooms, which allows a very advantageous business for sites that pay for visits to their sites on the internet. Also other Chatbots, were dedicated to stealing user data, sending spam, distributing malware, and creating phishing attacks. To avoid the large number of these types of Bots, tests such as CAPTCHA were created, however, they managed to pass those measures.

One of the techniques they developed to block this type of software was the analysis and classification using Conversation Log, this test is based on the Turing test to analyze whether it is a bot or a human. In addition, shared URLs, response type and conversations are analyzed, and different behaviors were found in the bots, for example: Newspaper Bot, Bot Responder, Replay Bots, and Replay-Responder Bots.

Tay’s Case: Racist and Xenophobic Bot

In April 2016, Microsoft released a Chatbot named Tay on the Twitter platform, this bot was programmed to behave like a 16-year-old girl, able to learn natural language and be able to chat with other users. The more users talked to Tay, the more the ability to learn about conversations was.

However, within 24 hours of being published, Tay posted hate messages to Mexicans, Jews, feminists, love for Hittler, support for genocide and even mentioned hating all people and in Tay’s last tweet he said, “be tired of humanity.” Users found certain fleas that allowed the bot to learn these kinds of behaviors and there was also omission from Microsoft to filter these types of attacks. This event is a great example of how susceptible artificial intelligence can be fueled by harmful people and incentivizes to create a global discussion about ethics in intelligence artificial y como debe de ser considerada por los desarrolladores de este tipo de herramientas.

Al final, Microsoft decidió retirar el bot mientras realizaba modificaciones, pero Tay fue apagada y retirada definitivamente y sirvió de ejemplo del aprendizaje del Machine Learning, incorporado en el Microsoft Bot Framework lanzado el mes siguiente.

With artificial intelligence, we are summoning the demon. You know all those stories where there’s the guy with the pentagram and the holy water and he’s like, yeah, he’s sure he can control the demon? Doesn’t work out. Elon Musk

I’m really optimistic. Technology can always be used for good and bad, and you need to be careful about how you build it, and what you build, and how it’s going to be used. But people are arguing for slowing down the process of building AI — I just find that really questionable. I have a hard time wrapping my head around that. Respuesta by Mark Zuckerberg to Elon Musk about his statements of fear of artificial intelligence. 2017

A. Accessible Advantages of Use : The use of ChatBots allows them to be used via text-voice and their response is in the same way considered a universal interface, which makes it easier for them to be accessed by people with different disabilities.

B. Allows you to know the user : Using Chatbots, allow you to send information about the most searched words by the user, types of purchase, behavior, loyalty and allows you to better train the bot to improve responses.

C. Availability: A service such as ChatBots, allow the user to have access to information 24/7, therefore, an always-available infrastructure is required.

D. Does not require human operators: There is no need for the intervention of a being human for each chat session, but sometimes, it is necessary for the Chatbot to be trained and have a human controller to provide the answers it fails to understand. The cost of a virtual agent, according to research published by Forrester, costs about $1 or less, while a human agent costs $5 or more.

E. Quick Information : Users require quick and accurate answers, if they leave the platform. This ease, is the big advantage of Chatbots because you can have open “N” number of sessions.

F. No need to install new apps: Due to the limitations of mobile devices (RAM, battery and storage), as well as the fact of switching between apps, it has been proven that users do not want new apps. According to Forrester research, it estimates that 80% of users use their time in just 5 apps and are mostly messaging applications. While mobile apps are difficult to maintain, distribute and create.

G. Conversations as an interface : Chatbots are the new era of applications because they can work on any device, the user’s way to enter data and obtain data is through a conversation. Conversation is considered more natural than clicking buttons, and they can respond with videos, images and audios. Interactions with text are fast, fun, flexible, intimate and more descriptive than other user interfaces.

Current Status We are currently in 4 new paradigms of the era of internet and mobile devices

  • Conversations are the new user interface.
  • Bots are the new applications.
  • Artificial intelligence is the new protocol.
  • The applications of messaging are the new search engine.

The current growth of the platform and development of bots is due to the use of messaging tools on mobile devices, computers and video game consoles and smart speakers such as Alexa. Messaging has also become the most used internet service, even beyond social networks.

Messaging applications concentrate the largest number of active users on the internet, for example Whatsapp has 1 billion Users, Facebook Messenger with 800 million Users, Skype with 300 million users and Snapchat with 100 million Users, to take advantage of the fact that users spend more time using their applications these companies have invested in the field of Chatbots and artificial intelligence.

In the table above, we can see how in the first quarter of 2015, the four largest messaging applications increased the number of active users.

Currently, the tools allow you to offer two types of Chatbots depending on the purpose of the developer:

1. Artificial intelligence-based features:

  • Understand and process the NPL language
  • They interact in a human form.
  • Able to improve responses, depending on training.
  • Able to offer various customer services.

2. Rules-based functions:

  • No artificial intelligence required
  • Trained to do only one thing
  • They are as smart as they are programmed
  • Communicate in a structured way
  • Command-oriented.

Google, Facebook, Microsoft and Amazon have started the race for this new era of communication with the acquisition of Startups specializing in speech recognition, pattern recognition, imágenes y desarrollo de inteligencia artificial.

Referencias Abu Shawar, B., & Atwell, E. (2007). Chatbots: are they really useful? LDV-Forum: Zeitschrift für Computerlinguistik und Sprachtechnologie, 22(1), 29–49. https://doi.org/

Hirschberg, J., & Manning, C. D. (2015). Advances in natural language processing. Science, 349(6245), 261–266. https://doi.org/10.1126/science.aaa8685

Watters, A. (2016). Clippy and the History of the Future of Educational Chatbots. Hack Education Weekly News. Recuperado a partir de http://hackeducation.com/2016/09/14/chatbot

Aron, J. (2011). How innovative is Apple’s new voice assistant, Siri? New Scientist, 212(2836), 24. https://doi.org/10.1016/S0262-4079(11)62647-X

Jo-Foley, M. (2013). «Cortana»: More on Microsoft’s next-generation personal assistant. ZDNet}. Recuperado a partir de http://www.zdnet.com/cortana-more-on-microsofts-next-generation-personal-assistant-7000020602/

Gallagher, S. (2012). How Google and Microsoft taught search to «understand» the Web. Ars Technica. Recuperado a partir de http://arstechnica.com/information-technology/2012/06/inside-the-architecture-of-googles-knowledge-graph-and-microsofts-satori/

High, R. (2012). The Era of Cognitive Systems: An Inside Look at IBM Watson and How it Works. International Business Machines Corporation, 1(1), 1–14. Recuperado a partir de http://www.redbooks.ibm.com/redpapers/pdfs/redp4955.pdf

Allison, D. (2012). Chatbots in the library: is it time? Library Hi Tech, 30(1), 95–107. https://doi.org/10.1108/07378831211213238

Wilks, Y. (2005). Artificial companions. Interdisciplinary Science Reviews, 30(2), 145–152. https://doi.org/10.1179/030801805X25945

Milani, R. V., & Franklin, N. C. (2017). The Role of Technology in Healthy Living Medicine. Progress in Cardiovascular Diseases, 59 (5), 487—491. https://doi.org/10.1016/j.pcad.2017.02.001

IBM. (2017). Watson for Oncology. Retrieved March 31, 2017, starting at https://www.ibm.com/watson/health/oncology-and-genomics/oncology/

Jenkins, M.-C., Churchill, R., Cox, S., & Smith, D. (2007). Analysis of User Interaction with Service Oriented Chatbot Systems. Human-Computer Interaction, 4552, 76—83. https://doi.org/10.1007/978-3-540-73110-8_9

Lu, C., Chiou, G., Day, M., & Ong, C. (2006). Using instant messaging to provide an intelligent learning environment. Intelligent Tutoring Systems, 575—583. Retrieved from http://www.springerlink.com/index/y511878064h24318.pdf

Microsoft. (s. f.). Skype BOT Available. Retrieved April 1, 2017, starting at https://support.skype.com/en/faq/FA34655/what-skype-bots-are-available

List, B. (2017). BOT List. Retrieved April 1, 2017, starting from https://botlist.co

Gianvecchio, S., Xie, M., Wu, Z., & Wang, H. (2011). Humans and bots in internet chat: Measurement, analysis, and automated classification. IEEE/ACM Transactions on Networking, 19(5), 1557–1571. https://doi.org/10.1109/TNET.2011.2126591

Anonymous. (2015). Ethics of artificial intelligence. Nature, 521(7553), 415–418. https://doi.org/10.1038/521415a

Beres, D. (2016). Why Microsoft’s Racist Chat Bot Catastrophe Was Kind Of A Good Thing. Recuperado 1 de abril de 2017, a partir de http://www.huffingtonpost.com/entry/microsoft-tay-racist_us_56f556e5e4b0a3721819bd15

Forrester Research, I. (2009). It’s Time To Give Virtual Agents Another Look. Online. Recuperado a partir de Forrester Research, I. (2009). It’s Time To Give Virtual Agents Another Look. Online.

Perez-Marin, D. (2011). Conversational Agents and Natural Language Interaction. Conversational Agents and Natural Language Interaction: Techniques and Effective Practices. Recuperado a partir de http://www.scopus.com/inward/record.url?eid=2-s2.0-84867490969&partnerID=tZOtx3y1

Mauldin, M. L. (1994). ChatterBots, TinyMuds, and the Turing test: Entering the Loebner Prize competition. In Proceedings of the Twelfth National Conference on Artificial intelligence (vol. 1) (pp. 16–21). Seattle, WA: American Association for Artificial Intelligence.

Anderson, S. L. (2008). Asimov’s «Three Laws of Robotics» and machine metaethics. AI and Society, 22(4), 477–493. https://doi.org/10.1007/s00146-007-0094-5

Ebling, M. R. (2016). Can Cognitive Assistants Disappear?. IEEE Pervasive Computing, 15(3), 4–6.

Modha, D. S., Ananthanarayanan, R., Esser, S. K., Ndirango, A., Sherbondy, A. J., & Singh, R. (2011). Cognitive computing. Communications of the ACM, 54(8), 62–71.

LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436–444.

Sebag, M. (2014). A tour of machine learning: An AI perspective. Ai Communications, 27(1),11–23.