"Human language is the new user interface, Bots are the new apps."(Satya Nadella )
"Chatbots are the universal user interface" (Dan Grover).
"Bots will not replace apps. Better apps will replace apps" (Dan Grover).
Apps
The term "App" (short for "application") was coined by Apple in 2008 to refer to small programs that ran on smartphones and allowed the user to perform specific tasks. Today, this technology has been consolidated. There are millions of Apps available, having also infiltrated tablets, personal computers, watches and TV sets. Apps have become popular because they are useful, easy to install and remove.
The problem with Apps is:
There are a huge number of Apps for different services. Many Apps are similar.
They are disconnected from each other. It is fragmented software, not integrated.
Each App has its own graphical user interface (GUI). There is no common communication system.
They are developed with different tools and languages.
GUIs were born in the 1980s by Xerox, Apple and Microsoft as a way to facilitate the use of applications through the desktop metaphor (icons, buttons, drop-down menus, toolbars and other graphical elements). GUIs avoided having to learn the different commands to access system functions.
The current trend is to replace application-specific GUIs with a common conversational system in natural language. It is about moving from GUIs to CUIs (Conversational User Interface).
This new approach involves the use of artificial intelligence (AI) techniques. AI is the intelligence exhibited by machines programmed with agents capable of perceiving the environment and making appropriate decisions, simulating human intelligence. In the case of the natural language interface, the system must analyze each sentence or message from the user, deduce what service is required, access the corresponding software and issue a response also in natural language.
The new era of technology is the era of the conversational interface, the advantages of which are:
It is a natural, simple and intuitive form of interaction.
Conversations can do things that traditional GUIs can't do. They are more flexible because natural language is flexible.
The interface can be in the form of text or voice messages, both input and output.
The user does not need to know which particular App to use to satisfy his requirements.
System responses are indistinguishable from human responses.
Text has many advantages. It is asynchronous. It is flexible. It can be efficiently indexed and searched, translated, compared, grouped, corrected, summarized, filtered by algorithms, group edited, shared, quoted, etc.
Bots and chatbots
AI, after a long period of lethargy, is getting a big boost thanks to the introduction of chatbots, intelligent systems that can converse in natural language with cell phone users to perform different services. The idea is to harmonize and synchronize the power of natural language and AI. This is what is called "conversational computing".
Nowadays we talk about Bots in general, chatbot being a type of Bot. The term "Bot" is an apheresis of "robot" and is the software version of it. Bots are applications that perform functions autonomously thanks to the AI incorporated in them.
The advantages of Bots are:
They provide automated services by exchanging messages. For example, request personalized information (news, weather, etc., even with graphic output), book hotels, make transfers, order food, translate a text, send an e-mail, order a cab, book a trip, make a phone call, filter calls and advertising messages, organize meetings, manage our calendar, remember dates, save notes, play music, etc.
It is not necessary to download them explicitly, as is the case with Apps, as they reside in the cloud. This reduces the time and effort involved in searching for the App that performs the function we need. A unified communication system is achieved where the Bots are hidden from the user. The user is only interested in the functionality, not the applications involved. Also the applications are updated automatically.
In early computers, computing started out as batch. With the advent of personal computers, computing became interactive. This interaction was first done with commands and then with graphical interfaces. With the advent of the Internet, applications ceased to reside on client computers, and only a web browser was sufficient as an interface. With Bots it is the same, they do not reside on smartphones.
They are activated by a user's request, but they can also be proactive: they act autonomously when certain circumstances occur or they detect needs to be covered. They learn from our habits and our behaviors and anticipate our requirements by showing us the things we want and need. For example, they recommend us books, videos and movies, the latest news in the fields we are interested in, recommend insurance, etc. The more the Bots are used, the more they learn and the better cognitive advisor we will have, and the more they can anticipate the user's needs.
Bots are interacted with in the same way as people. They are even treated as if they were people. For example, in the Skype App, the "add contact" option shows two alternatives: person or Bot.
It is not necessary to learn how each Bot works.
It doesn't matter which Bot you are talking to. The output will always be more or less in the same format.
There are many types of Bots:
Conversational Bots (chatbots).
Virtual assistant bots.
Bots that interact with other Bots. For example, collaborative Bots.
Bots that supervise other Bots (hierarchies of Bots).
Bots that connect devices, as in the Internet of Things (IoT).
Bots for business automation, administration, finance, etc.
The new paradigm of the digital world is a conversational system based on a universe of Bots and that makes use of the most appropriate Bot at each moment to meet the user's requirements. It is about interacting in a more natural and effective way with computers, which will make our life easier.
The new revolution, the new frontier, is that of Bots, although they are not yet popular because the technology is in its infancy.
MENTAL, a Language for Apps and Bots
Intelligence vs. Consciousness
Indeed, after years of inactivity, AI is getting a boost thanks to cell phones and Bots. The problem with Bots is that they require AI languages. And today, with the tools available, it is difficult to develop AI and conversational applications.
With MENTAL we have the simplest language possible and at the same time the most powerful language possible, the universal language, a language that allows the development of Apps and Bots. It is a multi-paradigm programming language and an AI language, simple but powerful, more powerful than Lisp and Prolog, as these languages are monoparadigm (functional and logical, respectively). It allows qualitative and quantitative automatic reasoning, fuzzy or imprecise.
With the archetypes of consciousness, automatic reasoning can be implemented. Intelligence is derived from the archetypes of consciousness. Consciousness is on a higher plane than intelligence; therefore, from the archetypes of consciousness everything is simpler.
So it's not about intelligence, it's about consciousness. And the ultimate approximation to consciousness is the primary archetypes. There are many types of intelligence, but there are only a few archetypes of consciousness.
Bots vs. Apps
Bots are types of Apps. It makes no sense to compare Bots with Apps, as it makes no sense to compare cars with transports: the car is a type of transport, and Bots are types of Apps. Therefore, Bots are not going to replace Apps as has been said. Apps will have more and more intelligence, with automatic functionalities.
With MENTAL, the development of Bots is essentially no different from Apps. And this is thanks to generic expressions, which allow to easily implement automatic functionalities. It is enough to program actions that are triggered when certain events occur in the abstract space. Generic expressions play an essential role in the integration of programming paradigms and AI functionalities.
Conversational system
Instead of having the Bots interact directly with the user in natural language, it is simpler to have a single conversational system. The process would be as follows:
The user input (natural language text) is transformed into a formal sentence S, following the MENTAL encoding.
The system analyzes S and deduces the Bot B that has to provide the service.
The Bot B analyzes S, processes it and generates a response T in formalized language (MENTAL).
The conversational system transforms T into a natural language sentence and sends it to the user.
In this way, the responses of the Bots have a uniform structure, and there is only one communication system between the user and the Bots, a system analogous to the browser.
Conclusiones
Every few years a paradigm shift occurs in the digital world. The currently emerging paradigm is the conversational interface and Bots, AI applications. But the universal and definitive paradigm lies in the archetypes of consciousness and the language that interrelates them.
There are two key elements in the new digital revolution: language and AI. Both aspects constitute the essence of MENTAL. This universal language is the tool that underpins and integrates all applications (Apps and Bots).
Addenda
Tech companies and Bots
Technology companies already have virtual assistants, Bots and tools for creating Bots:
Microsoft.
Microsoft's vision and bet is "Conversations as a Platform" (CaaP). Bots are the new applications. The goal is to create a network of Bots that can interact with the user in an intelligent way. Microsoft has created an AI division with the goal of "democratizing" AI. It has created Bot Framework, a platform for developing Bots that works with any programming language.
Cortana is an intelligent virtual assistant, Microsoft's main Bot. It recognizes natural language and uses Bing (Microsoft's search engine). It keeps a "notebook" where it stores all user information. Microsoft wants Cortana to be the backbone of everything.
Skype (a Microsoft company) has Bots developed by Microsoft. Skype Bot SDK is for developing Bots for Skype.
Tay was a failed Bot from Microsoft. It was designed to hold conversations on the social network Twitter and targeted a young audience. It was created as an experiment in computer-human interaction. It was only active for 16 hours because of its racist and xenophobic comments.
Apple.
Siri is a virtual assistant with voice recognition and voice response. It adapts to the user's needs, makes recommendations and performs various tasks such as reminders, contacts, notes, web browsing, maps, music, weather information, stock exchange, etc. Supports multiple languages.
Google.
Google Assistant is a conversational virtual assistant. It is an enhancement of Google Now and an extension of "OK Google". Google Now knows relevant information about the user: where they work, their hangouts, information they are interested in, sports they like, etc." OK Google" does voice searches, controls devices, sends messages, etc. On smartphones it is the Google Allo App.
Facebook.
This company has turned in earnest to the world of Bots. There are more than 10,000 developers working on Facebook Messenger ChatBots. According to Mark Zuckerberg, Messenger is not a messaging app, but a platform. The idea is for all businesses to be on Facebook and all integrated and communicating through the Facebook Messenger platform.
Amazon.
Echo is a voice-controlled speaker. Echo connects to Alexa Voice Service, a virtual assistant that is activated with the word "Alexa." All you have to do is make a spoken request: request music, request information, send emails, receive your shopping list, etc. It aims to make everyday life at home easier. Amazon has its own service: Amazon Music Unlimited. Echo has 7 microphones so you can hear from anywhere in a room, and produces immersive 360° sound.
Alexa goes beyond Siri and Cortana. It can be placed anywhere in the house, and is always ready to take commands. It's not locked to a phone, computer or tablet. No need to get close to it or press any buttons.
Others.
Slack is a teamwork tool that resides in the cloud offering various services. Slackbot is a Bot that helps organize meetings and performs task reminders. It is activated with "Hi, Slackbot here!!!".
Telegram is a veteran in the use of Bots. It has a Bot Platform.
Line also has a Bot development platform.
The WeChat revolution
In the West, we use an app for everything. In China, on the other hand, they use an integrated application (or platform or application of applications, even an operating system): WeChat (called "Weixin" in Chinese), developed by the Chinese company Tencent. It is the most used application in China and the benchmark to follow. It is not the "Chinese WhatsApp", as it is sometimes called. It is the most impressive phenomenon in the mobile world. Initially it was a messaging application, but it evolved to integrate many services:
Instant messaging. Type WhatsApp.
Moments. Sharing of text, photos, videos and links. Social network type Facebook.
Contact with physically close people. Type Tinder.
Shaking the phone, we contact people who are also shaking the phone at that moment.
Message in a bottle. Write a text and throw it into the metaphorical ocean of cyberspace where anyone can pick up the virtual bottle, read the message and reply or float it back.
Phone calls to landlines and mobiles. Video calls. Type Skype.
Organize group work. Type Slack.
Payments. Pay bills, send money, etc. Type PayPal.
Travel. Type Trip Advisor.
Marketplace. Application development platform for commercial relations. Content dissemination. Automation flows. Electronic business card exchange.
Various services: order a cab or a private vehicle, make a doctor's appointment, book train or plane tickets, book a hotel room, buy tickets for shows, purchase goods and services, buy and sell houses, etc.
The Turing test
The Turing test was proposed in 1950 by Alan Turing in his paper "Computing Machinery and Intelligence," which begins with these words, "I propose to consider the question, 'Can machines think?'" Turing considered that "thinking" was something indefinable because it was an internal process, and that it would be simpler to evaluate its external behavior. So he changed the initial question to "Can machines do the same as humans as thinking beings?". He defined or associated intelligence with the indistinguishability between human and machine.
The test involves three players: two humans and a computer. The player C (human) types questions on a computer terminal and receives answers on the same terminal from player A and B. The game is that C must find out who is the human and who is the computer. If C fails to find out, the computer is considered to have intelligence; it has passed the Turing test. The test does not consider the machine's knowledge in terms of its ability to answer questions correctly, but only its ability to generate answers similar to those that a human would give.
Turing was inspired by a social game that was fashionable at the time, the imitation game. It consisted of three players: A (a man), B (a woman) and C (of either sex), who is the interrogator. C only has contact with A and B through written notes. The object of the game is for C to find out who is the man and who is the woman. The player A tries to fool C with his answers (to make him think he is of the opposite sex). Player B is sincere in his answers.
In 1980, John Searle proposed his famous Chinese room thought experiment. According to Searle, the Turing test is meaningless because machines only manipulate symbols, they do not "understand" what they are doing, so they have no intelligence.
Eliza
Eliza was the first chatbot in history. It was created by Joseph Weizenbaum in 1966 to try to pass or challenge the Turing test. It was a program that mimicked a psychotherapist. It analyzed the user's words in search of keywords. If it found any, a rule transformed the original message and sent it to the user in response. If it did not find any keywords, it gave a generic response or repeated a previous comment.
The program was too basic (it consisted of only 200 lines of code) and could not go beyond a small conversation. However, it seems that it was able to fool some people into thinking they were talking to a real person.
Bibliography
Hawkins, Jeff; Blakeslee, Sandra. On Intelligence. Holt Paperbacks, 2005.
Norvig, Peter. Paradigms of Artificial Intelligence Programming. Case Studies in Common Lisp. Morgan Kaufmann, 1991.
