Recently, user interest in communicating on social networks and using mobile applications has been fading somewhat, and chatbots are becoming more and more popular, capable of giving out exactly the one that interests the user from a huge amount of information. The popularity of chatbots is primarily associated with the development of instant messengers, in which users spend more time than in social networks and applications.
The technology behind the advent of chatbots is natural language processing (NLP). NLP-based chatbots have passed the peak of excitement cycle expectations and are moving toward enlightenment and productivity. NLP provides chatbots with the ability to understand and interpret a user in their native language to offer a richer experience of communication.
What is a chatbot?
Chatbots are intelligent dialogue computer programs that simulate human communication in its natural form. The chatbot can process user input and output. Typically, chatbots accept natural language text as input, and the output should be the most relevant output for the sentence entered by the user. Thus, chatbots are an automated dialog system that can simultaneously serve thousands of potential users.
Chatbots have been around for more than half a century. The first programs, which today can be called chat bots, appeared in the middle of the last century. The introduction of the term "chatbot" is attributed to developer Michael Moulding, who in 1994 created the communication program Julia and called it a chatbot. The program was uncomplicated, but allowed to maintain a more or less meaningful dialogue and supported the functions of self-study.
Evolution of chatbots
The evolution of chatbots took more than 50 years. In 1950, computer scientist Alan Turing developed the theory that the human brain is a digital computational mechanism that over time is trained to become a universal machine. He is known for his pioneering work in artificial intelligence (AI) and cognitive science. In 1950, he created a Turing test to determine if a computer was thinking.
After Turing, researchers continued to work on what we know today as chatbots. They used a variety of technologies, such as natural language processing and artificial intelligence, to create the most artificial human experience possible.
The first generation of chatbots and natural language processing
At the beginning of their development, chatbots could not offer the benefits they offer today. Instead, the first-generation chatbots showed clear flaws because they didn't offer a good user experience. Firstly, at this stage they did not have real artificial intelligence capabilities, they offered scripts and robotic user interfaces. These rule-based chatbots worked acceptably for simple FAQ content, but even at this point a new horizon of functionality was opening up: chatbots could potentially do much more. Early versions were also saddled with long payback times - at least 9 to 12 months for assembly and deployment.
If you remember the first chatbots, then in 1964 Joseph Weizenbaum, a scientist at the Massachusetts Institute of Technology, began work on ELIZA. By 1966, ELIZA seemed to be talking to people. In fact, ELIZA mimicked the words of people who spoke to her. She replaced human words and used them in her responses, making it look like she was talking to them.
Nearly a decade later, Kenneth Colby developed the basic principles of ELIZA by creating PARRY, which employed a more conversational strategy. In 1973, there was a conversation between ELIZA and PARRY. These were early attempts to use natural language processing in chatbots.
As such, these first rule-based chatbots were primarily suited to the simple content of frequently asked questions. Despite the small feature set, the creation and implementation of the chatbot also took a long time. Moreover, because they were not based on artificial intelligence, they could not learn on their own, and constant manual training was required to teach the chatbot new tasks or questions.
Artificial intelligence comes to the fore
Since the first chatbots were not cost-effective, engineers found a more flexible approach to using artificial intelligence in the context of chatbots. The chatbot itself is nothing more than a programmed set of answers. Consequently, its ability to offer special benefits and exciting customer experience is more realistic when combined with AI. This allows the chatbot to reference historical data and formulate responses independently - and thus becomes conversational AI.
The first attempt to create artificial intelligence through human interaction was Barmaglot. It was created in 1988 by Rollo Carpenter. According to Carpenter, it was designed to "mimic natural human conversation in an interesting, entertaining and humorous manner".
In 1995, an artificial linguistic Internet computer entity, or ALICE, was released. In its original form, it used natural language processing. Since its introduction in 1995, ALICE has undergone several changes, switching to the language of artificial intelligence. The Turing test, unfortunately, did not pass, but was repeatedly recognized as the most "human" bot among the existing ones.
Consumers played a crucial role in this development. Customer service is a top priority for any business, and today customers expect nothing more than high-quality and personalized user experience. After all, without satisfied customers, the company will not be able to survive in the long term.
In the digital age, this could be a challenge for businesses. There are always new channels that need to be integrated to offer customers the optimal quality of service. In addition, companies are increasingly expected to communicate with their customers in the same way as with colleagues, friends or family. The rules-based chatbot doesn't understand this context. Therefore, it is clear that chatbots using AI will be successful in the future.
Chatbots in the recent past
In the early 21st century, chatbot technology struggled to provide a reliable and complex experience. Some solutions took too long to implement. Others were less reliable, spewing inappropriate or offensive remarks.
In 2001, the Smarterchild bot appeared, which proved that people really like to communicate with a smart machine that has its own character and can give good advice.
Later, in 2006, IBM developed Watson. This chatbot is capable of processing natural language and learning through communication. It was from him that large companies became interested in chatbot services, as they realized that they could replace most operators.
2007 was the beginning of the development of Apple's cloud personal assistant, which supports many languages and has an API (application programming interface) for integration with other applications.
Siri, which made chatbots part of our daily lives, was developed by Apple in 2011. It is able to answer questions and carry out various orders on the network. For example, if you say "Siri, turn on the music" and name the artist, then in a couple of seconds you will already listen to what you asked.
Google Now - a personalized assistant from Google appeared in 2012. This smart chatbot was created because Android users could not use Siri. A few years later, he grew up in a full-fledged Google Assistant.
The famous voice assistants Alexa from Amazon and Cortana from Microsoft appeared in 2015. They have further integrated chatbots into everyday life. Smart programs have learned to easily recognize human speech, learn, answer all kinds of questions, respond to voice commands, order goods at home, and much more.
In the same year, Telegram opened an application for developers to create and deploy bots. Following this, Facebook opened its messaging platform Facebook Messenger to develop bots attracted the attention of users as well as developers. Now almost every popular platform, including Whatsapp, Slack, We Chat, IMO, offers the ability to create a chatbot.
In 2016, the network was conquered by the famous Tay - a self-learning bot from Microsoft, which was supposed to adopt the manner of communication of teenagers. Known for learning from the answers of people, as a result of which in less than a day he turned from a cute robot into a paranoid racist. Microsoft had to close its project until they figured out how to deal with provocative user messages.
With the development of technology happening daily, who knows what the next step in the evolution of chatbots will be. Perhaps people will report symptoms to the nurse's chatbot. Or students can have metaphysical discussions with philosophical ones.