Many people fear that artificial intelligence will get out of hand. With our story, we want to clear up rumours and show why these fears are unfounded.
It began as a normal day: In the morning I got on the train, rode to the office and sat down at my desk to start my working day. I have been working in this company’s customer service department for over two years now, and every day I took care of the numerous customer inquiries. Over the weekend, I have completely forgotten that from this week on, someone new would join our team, namely the AI robot. He seemed to be doing his job properly and yet I found myself looking at him a little more critically. You never know what such an Artificial Intelligence is up to - maybe there is a spark of truth in all those scary science fiction movies where Artificial Intelligence gets out of control and brutally takes over the world?
In any case, our robot was only there to talk to our customers, to answer their questions and to remove possible worries. During a short coffee break, I sat down next to him and watched him communicate with the customers in the chat.
One strange thing immediately caught my eye: Without dot and comma, without the simplest of greetings, he threw clumsy statements at the customers. A problem with the configuration probably, I thought to myself. But after I told him that we must always respond politely to customer enquiries, his messages became even bolder to the point where he even insulted customers. When I took a closer look at his body, I realized that he had a loose screw and not just figuratively. With every passing second, his madness increased even more. He he turned his face expression dark, so that I could not assess him at all. His case was completely overheated, already smoking from the ventilation and he made frightening squeaking sounds while hammering aggressively on the keyboard. I had to act fast.
I knew that there was only one person who could help me: I quickly sprinted into the next room to fetch a colleague from the technical department. But when we returned to my workplace together, everything was upside down. The entire office was destroyed, paper flying around everywhere, all computers were swept from their places and laid on the floor in pieces. As if it had been driven by an evil force, the unbearably squeaking and smoking robot tore the cables out of the sockets and threw all the computer mice through the window.
When the robot reached for a keyboard to smash it on one of the remaining screens, my colleague rushed to him to rip it from his hands. But of course, no human being is too good for the cruel robot; without even hesitating, he took a firm grip on it and shoved it down the throat of my screaming colleague.
In the reverberation of the noise I was frightened out of my bed. Immediately, it became clear to me: my little scepticism about Artificial Intelligence turned into a powerful fantasy in my dreams. This was not really helpful, because my company was actually expecting a new AI robot and I was supposed to work with it.
I eventually pulled myself together and sat down at my workplace. The robot was already sitting on the place next to me, greeted me kindly and introduced himself. Do not let your nightmare fool you, I thought to myself, when the robot asked me a question: Initially, I had to show him how to answer customer inquiries. He seemed to understand my instructions very quickly; after all, he was supposed to answer a large part of customer inquiries independently without my help. Actually, my work was more fun than usual: As the bot was answering recurring customer questions automatically, I had more time and energy for individual cases. Because of the positive feedback from our customers, I felt certain that they also appreciated the quick and uncomplicated service.
No company that introduces Artificial Intelligence is likely to experience the destructive dimensions of this story. Nevertheless, a lot can go wrong, especially if you develop AI systems that have no control protocols for people (for example, for the service employee in customer service). A prominent example of this is the Chatbot Tay from Microsoft, which everyone could communicate with. What started as a fascinating project resulted in racist and sexist insults of the chatbot against the users. Although this is by no means as brutal as in the above example, it nevertheless proves that the skepticism of many people about Artificial Intelligence is not completely unfounded.
However, the misconduct of the Microsoft bot was due to the fact that Internet trolls deliberately used sexist and racist expressions to train the bot to do this. This is exactly what can happen when systems are developed that do not provide sufficient control protocols, or do not provide controls protocols at all.
The previously stated example is not the case with OMQ, as it is clearly defined by a service manager how the chatbot responds to what kind of request. AI learns that customers express themselves differently. In doing so, AI is able to associate questions and provide the right answers. Here, our AI works like an employee - similar to the smart robot in the second part of our story. The advantage with our products and especially with our chatbot is that an answer is defined only once. Afterwards, our Artificial Intelligence replies independently on all channels.
So you can definitely be all-clear: Artificial Intelligence is neither aggressive, brutal, nor otherwise dangerous. In order to prove this and to take other fears, we will publish more scary articles in the category ”Scary New World” in the future. So there will be exciting texts with creepy pictures and a lot of know-how about artificial intelligence in customer service regularly! If you don’t want to miss any more blog articles and stories, subscribe to our newsletter.