The Phenomenon of a Human-Like Chatbot and Why You Already Want One

10 Nov 2017
9 min
author avatar
Daria R.

Welcome to the new era. The era where stamps, landline phones, and handwriting have become relics. No wonder that we prefer a company of machines to people. We talk to them, share our secrets and emotions and sometimes even can’t understand who we are talking to, a human or a program. Frankly speaking, probably we don’t want to know. Today, we decided to shed some light on the phenomenon of the human-like AI chatbots, how they manage to fool us, and why our own bots have to be like “dudes” to become popular.

How do human-like chatbots trick us?

If you ask your friends to distinguish a real person from a bot, they would say that it is easy. However, the latest real-life events prove that it is not even close to a reality. There were so many cases when realistic chatbots weren’t identified as programs and even were asked for a date, but more on that later. So how do they fool us? Let’s take a look at the strategy bots lead to misguide us:

human chatbot

How bots trick us

Ask again

What happens if we haven’t heard a phrase? We ask again. So do they. They make pauses, say “Um” and “Huh” in the situations real people could feel confused or surprised. They type tree dots to intrigue you. It helps them capture your attention and continue the dialogue.

Start over

Every question you ask a bot is not a part of a dialogue. Bots don’t lead dialogues, they simply ask or answer questions. If you ask a bot about something you have already discussed again the bot simply won’t understand you or will give you another answer. They memorize only “useful” information, like your preferences and so on, but these human chatbots don’t need to keep in memory everything else. This is the reason why they clear the screen after each question, so you can’t check the exact answer of an interlocutor. And then you think that you could mix it up.


Need some help with your bot? Here's a clear guide how to build a chatbot for you!

Becoming a shrink

Not only they become smarter, but also give us a feeling of understanding and support. Human-like artificial intelligence bots can ask a user about their feelings, and what bothers them. Such analysis helps people understand the nature of their emotions better. Along with all that, customers feel the attention. They believe that interlocutor cares about them.

It has a character

Probably, this is the main thing that persuades us that we are talking to a person. Usually, the developers choose a very bright characteristic which, as we believe, can only have a real person. For instance, selfishness. If in a conversation someone talks about himself or herself all the time, it makes us believe that it is a human being. The same with weepers, scepticism, and sarcasm. It is strange, but we still believe that only people can have them. 

They can fight

We are used to thinking that all bots are so polite, they are well-bread because we taught them to be so. However, if a bot mimics a bad person, it’s almost impossible to expose it. Quarrels are irrational, they don’t care about grammar or repetitions. All that we feel is that the “person” we are arguing with is a stubborn hooligan but not a machine.

Here’s a story for you. One bot told everyone that he was a janitor and that he was afraid to lose his job because recently he messed up at work. Every time when someone said the word “job”, “work”, “boss” or whatever, the bot told about his “issues” at work. That’s how the program persuaded users that he was real.

Why users like human-like AI chatbots

There are several reasons why users prefer human-like AI chatbots to usual bots and sometimes even to their friends and family. They like to talk to a person more, but usually, our friends have little time for us, our parents are always busy or at work. And even if we go to visit someone, we have only a few hours. It’s not enough. And the fact that chatbots are always “here”, they always have time for us and are ready to listen to about all our problems. There are even more factors which influence our desire to talk to programs which act similar to us. Here are some of them.

ai human robot

Why your users need a human-like bot

They sound clever

And they really are. But only in their sphere. Nevertheless, the time of stupid chatbots is in the past and if we even can’t distinguish bots from people it’s a sign that they are definitely got smarter.

They don't judge

Unlike people. It is especially important if we deal with kids and teenagers. They are too sensitive to criticism and they very much need support and praise.

They won’t tell users' secrets

It matters to anyone. Of course, we understand that some data about users can be collected, like what they like doing, eating or listening. However, lots of people prefer sharing their emotions and problems with complete strangers. Such human-like chatbot can become your customer's stranger in a difficult situation. And they can be sure that the program won’t tell any secrets to their parents, girlfriend or colleagues.

They are efficient

Again, unlike people. A program cannot “forget” to perform a task for a user. It does everything straightforward. And your customers can give it as many tasks as they need.

Common people like smart bots for all that, they trust them more and give more information if they believe that this is a conversation with a human. So, if you already have a bot, check its effectiveness. Maybe it's time to make your bot more efficient? If you don't have a bot yet, read our tips to create a smart AI chatbot below. 

BTW, if you are an Android developer you may want to read our next article about the best Android frameworks for testing. Don't forget to subscribe to our blog and be the first to get our freshest articles!

How to make a human-like bot

You see how it works. People trust human-like bots, they share secrets with them, chat and even ask over. So, what does it take to create a human-like bot? Here we go:

ai and humans

How to create a realistic chatbot


Of course, a machine can’t have any. That’s why we teach them. Teach them to express emotions a regular person would have in a particular situation. The other task is to “identify” users’ feelings. And react in a proper way. Jokes is a very important part. Only the cleverest machines can not only tell a joke but to “understand” it and “laugh”.

Details matter

Remember what we said about irrelevant information? That’s right, most of the bots do not need it so they don’t memorize it. Your task is to make friends with a user, that’s why some details should be stored in memory. It helps build strong interaction with a user.


People are creative, even if they don’t have an answer on something they can make it up, they can change the subject of conversation or say something like “sorry, I have to go now, let’s discuss it next time” or so. Bots don’t do that if you haven’t taught them to. If your program can’t answer literally any question and gives the sentence we all hate “sorry, I don’t know”, it means you still have to work on it and your bot isn’t ready to be launched.


The easiest ways to make money with a bot are here. All the hints and tips from Cleveroad!

Talk like a human

It means that in some cases a user has to wait for half a minute for an answer and in some get it very fast one. Bots speech has to be alive and contain various interjection, emoji which correspond one or another “mood”. And don’t forget about jokes, it is one of the brightest abilities of a human which can persuade users that they are really talking to a person.

Applying all these tricks you can teach your bot “be a human”. All these techniques aren’t new and have found their realization in already existing bots. Let us take a look at the most successful chatbots which managed to fool users for a long time.

The most realistic chatbot

It’s been awhile since the first chatbot was launched. And People got used to talking to them. However, the new type of bots, human-like bots which are difficult to recognize have appeared not so long ago. One of the most famous is Amy.

Amy Amy is your personal assistant who schedules meetings for you. She is not only an efficient “worker” but also very pleasant to talk to. Maybe that’s why Amy was asked over for so many times (almost every month a bot got an invitation). In order not to make a person angry Amy gave a very honest answer. She said that she’s an artificial intelligence personal assistant and she cannot join a man in person. The developers who created Amy say that they decided to make their program behave like a real person to create an emotional connection between a user and the program. And it wasn’t an easy deal.

Humani Jessie

Jessie is a Facebook bot which is looking for a job, apartment and so on. Talking with Jessie you dive into “her life”. The program behaves as your friend who asks about your opinion about her life. She can tell you something like “Oh, I have to make myself some coffee, I’ll be right back in a minute” or something like that to persuade you that you are dealing with an alive human being. Because robots don’t drink coffee, right? Jessie doesn't perform any specific tasks, however talking to her is a real entertainment.

If you want to know more about the future of chatbots and if they are really going to replace usual applications, watch our short video below.

Ross the lawyer

Ross is a program based on AI whose main task is to provide customers with legal advice. Ross was created by IBM company, it is very intelligent, it can quote the law and provide the correct answers to any legal query. Its developers plan to cover the law of every country in the world.

As a bonus, we would like to introduce you to Tay chatbot from Microsoft. Unlike all other bots we have named today, Tay became a joke and represents a failure anyone can meet if he or she would launch an unperfect bot. Tay was supposed to be a teenage girl, who talks about Taylor Swift and Miley Cyrus all the time. But she learnt too much from real people and all that stuff was too creepy. Only in 24 hours, Tay started saying that Hitler did nothing wrong, Bush did 9/11 and asked men to have sex with her. Right now, the service is unavailable. We hope that Microsoft will fix this “child from hell”.

There is no need to tell you how effective and cost-efficient chatbots can be. But everything moves on, so do the technologies. Chatbots evolve and take different forms. If you have one, you should think about its perfection. For instance, to make it more like a human. If you think that this time has come, contact us and we’ll do our best to make your bot behave like a next door buddy!

Rate this article!

An imageAn image
An imageAn image
An imageAn image
An imageAn image
An imageAn image
Love it!
(891 ratings, average: 4.53 out of 5)
Back to top