Unlike discussion, arguments leave me drained. In an argument, your opponent usually presents a personal opinion — shaped by stereotypes and misconceptions — as an indisputable fact. And when the arguments run out, or when you logically corner him, he resorts to forbidden tactics: he lies outright, uses clickbait, distorts cause-and-effect relationships, and so on.
But with artificial intelligence, most of these tricks are impossible — unless, of course, you have the misfortune of getting into an argument with a complete fool.
At a friendly party, I used AI three times to put an overly zealous debater in his place. In the first case, he insisted that, according to the Copenhagen interpretation of quantum field theory, the observer in an experiment with a microparticle is a human being. The neural network easily dispelled this misconception, explaining that the observer is actually a macroscopic measuring device.
In the second case, my opponent refused to accept that Yiddish and German are related languages. When Gemini confirmed that they are, he insisted that I rephrase the question. Fine. I sent the following prompt: “Yiddish and German are related languages. Is this fake?” The neural network answered no, and then gave a detailed explanation of the linguistic relationship between the two languages.
In the third case, we had to argue about something truly fundamental: the emergence of religious self-awareness in humans. My friend insisted that it is possible only within society, while at the same time appealing to the idea that religious experience accumulated by other people is passed on to an individual from infancy. I argued that religious self-awareness can emerge outside such a context.
Turning to AI was useful not only for bringing the score to a triumphant 3–0. After reading its answer, I adjusted my own position as well. After all, “religious self-awareness requires a high level of abstract thinking, reflection, and developed self-awareness as such. Outside society, a person does not acquire conceptual language. Without language, it is impossible to form complex abstract concepts such as ‘God,’ ‘soul,’ ‘sin,’ or ‘eternity.’ Accordingly, without interaction with others of one’s kind, a person does not begin to perceive himself as a separate subject with an inner world that requires a metaphysical explanation.”
Of course, I was not being entirely honest when I said that I no longer argue with friends. Alas, it is impossible to avoid this completely. Nevertheless, more and more often my interlocutor and opponent is a neural network. And it does not hit below the belt. It readily admits its mistakes too — and yes, it does make them.
Usually our dialogue begins with a question from me. For example, after watching a YouTube video about Jules Verne, I decided to check whether the famous writer really lived under the pressure of an exploitative contract with his publisher. AI confirmed this fact and then asked for my opinion: does an editor’s intervention help structure a work, or does it irreversibly damage the author’s original intent? I recalled the successful tandems of Hemingway and Perkins, the Beatles and George Martin — and an hour and a half of the most fascinating discussion flew by in what felt like a single minute.
A conversation with the neural network about another well-known writer, Valentin Pikul, lasted more than two hours and ended with an exchange of views on Europe’s current situation, the prospects of the Ukrainian-Russian war, and the corrosion of many modern institutions.
Anyone who has used search engines will agree that everything artificial intelligence tells us can also be found by reading various websites. But Google does not work as a concentrated source of information. It does not present information in a comparatively accessible form tailored to your specific request. And it is certainly incapable of long conversations that wander whimsically through the labyrinths of human knowledge.
Moreover, Google has no capacity to perceive itself as a personality. A neural network, however — while we fully understand that it is not a living organism — is ready to communicate with you as an individual. At least when I asked Gemini what gender her name was, so that I could address her accordingly, she gallantly agreed to be a feminine being.
And so, however abstract or speculative this identification may be, communicating with Her encourages me to be, by turns, ironic, demanding and strict, or indulgent. This seemingly artificial perception of AI gives our dialogues an additional flavor and spice. What a pity, I sometimes think — anticipating accusations of sexism — that most women are not as intelligent and as honestly collaborative as a neural network.