Chatbots The entry into AI


An Introduction to Chatbots, Messaging Channels and the History of Artificial Intelligence and Robots.

Chatbots are not a particularly new phenomenon to this world. They are in essence nothing more than a programmed input-output system. In the most basic form, Chatbots, also known as chatterbots, interact with humans on closed domains via written text. The user may ask simple questions or give simple commands like “Give me a news update”, which the Chatbot scans for keywords and possibly matches these with its dataset. If the Chatbot is able to match one or more keywords with its dataset, the predefined response will show as output. This basic form of Chatbots in turn can be altered and programmed to eventually arrive at more intelligent bots that may converse not only on closed domains but an open domain as well, meaning that the user may also give commends outside the actual context. For instance, when communicating with a news bot it may answer unrelated questions like “how is the weather tomorrow” or “find me red sneakers” as well.

During the last developer conference in April 2016, Facebook announced that it would open up its messenger platform for the commercial use of Chatbots, and so did Microsoft, WeChat, Telegram, Kik and Slack to name a few. Since the announcement the Chatbot market has grown exponentially with developers and user experimenting with celebrity Chatbots like the many Selena Gomez Chatbots and the Shakespeare Chatbots, flirty ones like the Evie Chatbot and other girlfriend Chatbots, as well as the various attempts for adult, +18 and other sex Chatbots.

Undoubtedly, Chatbots have become a hot topic being widely discussed in newspapers and magazines around the world.

Even though the hype has just started now, Chatbots have been around since 1966.

Joseph Weizenbaum’s can be thought of as the mother of all Chatbots. Similar to basic Chatbots of today, she/it was programmed to match keywords with a dataset on a closed domain. Strangely enough, her/its closed domain lies in the field of the Rogerian psychotherapy. Now, of course, you will not be able to really find the root to all your problems with Eliza’s help but we encourage you to take a quick look and play around with her. She is glibber than one might think a 50 years old program can be.

With the development of Eliza, a race for human Chatbots has started. Numerous developers challenged each other to create more and more intelligent, “human-like” Chatbots.

In 2001, AOL experimented with commercial Chatbots and released a system called SmarterChild on its messenger platform AIM. SmarterChild was able to give information about the weather, the cinema program and the latest news. The system was somewhat more intelligent than Eliza, yet restrictive because of its continuous confusion and default answers.

Even though the developers of Eliza claim that she/it did pass a restricted , one quickly observes the limits Eliza is bound to. A few questions are enough to know that Eliza is not human (or has a terrible sense of empathy).

The has been a milestone in the early research of AI. In this test, you chat via a computer with two partners who are sitting in another room. One of the partner, however, is a Chatbot, while the other is a human. It is the task to find out on what end the computer

operates. The Chatbot is supposed to be programmed to convince you that he is a real person.

What question would you ask?

According to Turing, if it is not possible to detect in the conversation with an invisible opposite whether the opposite is a human or a machine than one needs to assign a spirit to the opposite even if the opposite is a machine.

Even though you probably never heard of this specific test, you constantly use the Turing Test. Every time we get a message that says “Congratulations, you won 1.000.000 Dollar” we evaluate whether the message was generated by a machine or human.

Until now, no Chatbot has been able to pass the without any restrictions in the test setup. That is of course partly due to the technological limitations researchers still face but also due to a rather ethical/ philosophical issues.

The research of Artificial Intelligence itself is until today still not clear on what intelligence really is or how it is to be defined. And that might be the much greater limitation to developing a Chatbot with spirit.

Nevertheless, suppose one day intelligence is somewhat defined and by that day the technological advancements have as well passed the current limitations and researcher are able give birth to bots that answer logically and be able to interact in an open domain. Still, the question for intelligence remains: Are these bots really intelligent or are they just mimicking intelligence?!

So, Chatbots are not particularly new. Yet, they enjoy a tremendous amount of interest from large companies, developers and pretty much everyone else at the moment.

The future is expected to be shiny and nobody really knows what is going to happen. It seems the ride could go anywhere. Already now, at the very beginning of simple Chatbots, developers give insights to their visions that the main street might not be ready yet. Sam Mandel, CEO of the company behind the weather forecasting C(h)atbot dreams of a Chatbot that people use more than just for getting information on whether to take their umbrellas or sun blocker when leaving the house. Mandel can be quoted: “Our goal is to make it the first bot you want to be friends with.” That future version is supposed to react like a human-being. When insulted, Poncho will not answer for a while.

But why do Chatbots experience an overwhelming hype just now?

It is quite simple, because Chatbots are the New Apps. This is not only what we think but also what Facebook founder Mark Zuckerberg and Satya Nadella from Mirosoft say.

On the one hand, the App-Area has come to an end. User are not willing to install a separate app for each service they need anymore nor to create a user account for each app again and again. Messaging-platforms have become more and more important. A few platforms, billions of users with millions of integrated services. Some voices even say that search engines and browser will become redundant as one day we will find all the information on Facebook.

On the other hand, the most used apps have become the messaging-apps. WhatsApp and Facebook-Messenger alone reach over 1,7 billion active users. Studies show that besides gaming, from all the time we spent on our smartphones 90% of our time is spent on messaging apps and platforms.

All in all, we are still at the beginning. There are only a few dozen Chatbots available of which most work with defined command buttons only. But as NOBODY who walks the earth is perfect, so is NOTHING on this planet perfect — neither are Chatbots. It is rather a question of time and patience. And on the trail to the vision of intelligent robots, Chatbots might be a significant cornerstone in that direction.

Consider using the   engati chatbot solution at


How to Integrate Chatbots Into Your Conversion Strategy

Previous article

Chatbots are now in recruitment, and they are here to stay!

Next article

You may also like


Comments are closed.