How Chat Bots Trick You
Chat bots have come a long way since the days of ELIZA, the psychotherapist that asked open-ended questions and provided very vague answers. Today’s chatbots are sophisticated enough to carry on a highly-engaging conversation with a human.
Technology has emerged to the extent where many organizations use it as the first point of contact. Using the pre-scripted responses to common queries, these chatbots could filter the incoming customer care issues into common and uncommon and direct the person accordingly. This allows an organization to maintain 24/7 initial customer care presence without the expenses of maintaining a call center and a CS agent.
However, the technology used in chatbots still cannot fool an actual human for long. In fact, people are often able to out the chatbot after the first few sentences. This is the most important challenge that the chatbot designers face. Since the current implementations of Natural Language Processing are not developed enough to stimulate the way a real person carries out a conversation, no chatbot could actually stimulate the experience of a “real” person. Thus, the developers often resort to following tactics to “trick” you into believing that you are not talking to a chatbot.
A Very Specific Character
People often identify other people through their character traits. A person could be dull, angry, selfish or very enthusiastic about a topic. These traits are relatively easy to implement because of their very limited range of requirements. Hence, instead of making a teenager interested in everything under the sun, it is easier to create and maintain a bot that mimics an engineer. The subject matter that the bot developers have to cover is narrower and easily categorized for instant access.
A very good strategy that compensates for the mistakes that the bot makes during the conversation is to make it a “foreign” entity. This is often the case when many bots appear to originate from non-English speaking regions. This way, the designers and developers have a plausible excuse for all the grammar and contextual mistakes that the bot makes during the course of the conversation.
Drag the Chat for Eternity
One of the time-proven tactics is that chatbots use is to drag the conversation using open-ended questions that invites the person to answer with a long sentence. This offers two significant advantages to the chatbot. The first advantage is the simulation of a meaningful conversation. The chatbot introduces new ideas based on the keywords it extracts from the replies. This leads the chatbot through new branches of the stored responses. The second advantage is the development in the large store of responses that the chatbot can use to improve their skills.
A very interesting trend in the development of chatbots is the idea of an argumentative personality. This ranges from grumpy to outright hostile persona. A fight is easy to sustain and the person fighting the chatbot supply most of the arguments. In fact, it is the person and not the chatbot that is responsible for sustaining the conversation.
Next time you are talking to a chatbot, keep an eye on how the chatbot developers have used one or more of these tactics to keep you interested in the conversation.