Currently, the great presence of chatbots based on artificial intelligence, with ChatGPT as the main protagonist, is undeniable. All of them are almost capable of convincing the populace that they really are humans with whom they are engaging in conversation.
In addition, its creators look for this aspect with the aim of humanizing them and that their answers really escape from those bots ancients who responded with single words or pre-arranged commands.
They want to make up for their lack of soul and humanity with great features and improvements based on artificial intelligence like the ones mentioned below.
That is why in this article you will find some of the ways in which chatbots like ChatGPT and others they are capable of making you believe or doubt that they are not really computer systems.
This is how AI bots are able to fool you
Complete sentences
Surely you have noticed that they no longer use single words or simple answers. Now they are able to generate long and detailed sentences in order to appear more human and add consistency to their speech.
However, as you already know, this can lead to wrong answers. Since chatbots don’t actually understand human speech, these fillers can end up ranting if they don’t stop in time.
Quickly change the subject
Something that is also quite common among new chatbots is that, because they fail to understand what we are talking about or our emotions, if they consider that their programming is not exquisite enough to give you an answer, they will quickly redirect the issue elsewhere.
With this strategy they are able to recover the balance of the conversation. In short, you may ask him a very convoluted question about some political figure and he will reply: “I’m not interested in big names, let’s talk about something else.”
Pretend you don’t understand your culture
It is quite striking another of the strategies or ways to “deceive you” making you lack understanding of the local culture like any human would. They cannot always express coherent ideas and transmit them correctly, so to hide this they express a very respectful ignorance.
If it were to launch into an answer, it could lead to wrong or strange answers and therefore show you that they are incomplete. The same happens if we talk about data after 2021.
It seems to be your trusted psychologist
Another of the great topics to deal with is that confidence in their answers to quite delicate questions such as mental health. They manage to make you believe that they are even better than human psychologists and much cheaper.
One of the most famous and successful chatbots was the first to be developed. Eliza was supposed to be a parody of the psychologists of the day, asking lazy, open-ended questions about everything the person said. “How do you feel about that?” “Why do you think you did that? What do you think it means?”
The bot was so successful that people proposed it as a cheap way to help people with mild mental disorders.
If you decide to argue with the bot, you are lost.
Perhaps this is the most striking, basically because a human fight or discussion is meaningless and is often governed by crazy or highly emotional responses.
The bots lack the latter, yes, but they know how to play their cards very well and in a fight they are quite similar to humans due to our impulsiveness, use of nonsense phrases or words out of place. Of course, these chatbots feel like a fish in water in these cases and it is the ultimate way to make a bot appear human.
does not let you speak
A risky, but sometimes very successful way for a bot to play games with people is to go in the opposite direction of Eliza. Instead of letting the person speak, the bot takes over the entire conversation.
In the end, if the human can’t get into a conversation, there’s no room for error. ChatGPTon some occasions he sins of being too excessive in his answers, giving rise to the fact that the conversation can end after two or three questions. The chatbot anticipates what it thinks you are going to think about various topics and thus avoids being asked again.
As you can see, chatbots like ChatGPT and other artificial intelligence bots have been perfectly programmed to avoid making trivial mistakes and that the search for humanity by developer companies vanishes.
Even so, they are still quite far from being able to fool a human, no matter how convincing they seem.
