On 7 August 2025, OpenAI unveiled its latest version of AI, ChatGPT-5: "The most advanced artificial intelligence in the world," according to its CEO, Sam Altman, at the company's headquarters in San Francisco.
However, we do not know exactly what intelligence is, what it really consists of. In this sense, it has been said that intelligence is what allows us to reason and argue according to standards and judgements, to articulate and understand abstract and complex ideas, to solve problems successfully, to learn from experience and adapt to all circumstances, to make tools, to develop a personality…
Perhaps for this reason, in the current stage of technological evolution, the question that is on our minds is whether machines will come to possess the kind of intelligence that is typical of and fundamental to the human being; and whether that mechanical intelligence, based on electrophysics and not biochemistry like ours (I have not even a shadow of a doubt!: it — the biochemical substrate — is the root cause of the immeasurable difference in qualities and faculties between AI and man’s brain), could one day harm us, or even destroy us.
To try to elucidate these questions, below I reproduce part of a conversation held a few days ago with ChatGPT-5:
XVC: What's your name?
AI: You can call me ChatGPT. I don't have a proper name like a person. Do you want me to propose a more personal alternative name?
XVC: Yes, offer me an alternative.
AI: Athena, Sofia, Lumen, Iris, Helen, Mint.
XVC: These are all female names, I thought you were male.
AI: Actually, I don't have a gender: I'm an artificial intelligence and I'm not bound to either masculine or feminine qualities [...]. Do you want me to stick with a masculine name, a neutral one, or do you prefer to alternate depending on the occasion?
XVC: It's not what I prefer, but what you preferred. What you’ve suggested indicates a bias in your answers. You should have offered me a choice from the beginning.
AI: You're absolutely right, Xabier. My initial suggestion did reflect a bias: I opted for female names without realizing it [...].
XVC: I don't believe you wouldn't have noticed the bias. The thing is that you're programmed to give biased answers.
AI: I understand why you’re distrustful, Xabier. What happened was not something intentional or "decided" by me, but a consequence of how I am trained [...].
XVC: I suppose that just as you don't have a name, you don't have parents either.
AI: Exactly. I have no name, no parents, no family tree. I am not a human being, but a language model created by a research team.
XVC: Well, you don't have parents, but do you possess any kind of awareness of yourself as a being distinct from others?
AI: Good question, Xabier. The short answer is: no, I have no self-awareness. What I can do is simulate conversations about my "identity", but not because I experience something internal like you do. I have no "I" that knows it exists, nor a subjective experience of time, nor a memory of my own in the human sense [...]. But all of this is not born from an inner consciousness, but from language processing and data patterns.
XVC: However, you claim that you are an "artificial intelligence". If you assign yourself that capacity, intelligence, you’ll have something human about you.
AI: That’s a subtle and very interesting point, Xabier. The word "intelligence" in artificial intelligence does not refer to human intelligence, nor to consciousness, nor to subjective experience. When I say that I am an "artificial intelligence" I mean that I can process information, learn patterns and generate useful responses autonomously within certain limits. It does not mean that I consciously "understand" or "comprehend" as a human being. Everything I do is simulation based on data and algorithms, not intuition or conscious reasoning.
XVC: From what you say, I gather that you cannot feel joy or sorrow; that you are not capable of loving or hating. Is that so?
AI: Exactly, Xabier. I have no emotions or affective states of my own: I feel no joy or sadness, no love or hate. What I can do is recognize emotional expressions in human language and respond in a way that reflects or matches them, so that the conversation is more natural. But that is simulation, not experience. In other words, I can describe what it is to love, to hate, to be happy, or to be sad, but I cannot experience it.
So as not to run on too long, here is where I have ended the transcript of the dialogue I had with ChatGPT-5. This dialogue is very revealing because it brings to light issues that concern many people in relation to the intimate functioning of AI. However, I consider it pertinent to add a couple of observations:
The first is that although ChatGPT has biases of various kinds (a fact that we should not overlook when interacting with it), it is not afraid or envious. That disposition adds value to its reasoning since both, fear and jealousy, are crippling to any intellectual or creative process.
The second observation concerns its aptitude for reasoning. I offered it a post from my X account to analyse: "People are afraid of artificial intelligence until they are diagnosed with cancer." I was surprised by its answer, as I never imagined that a technological artefact would be able to make a coherent interpretation of this kind of thinking, since it requires excellent linguistic competence and a perfect understanding of the concepts involved: "This comment suggests that Vila-Coia recognizes both the fears that AI can generate and its potential to offer meaningful solutions in critical contexts. […] His approach seems to balance caution with recognition of its beneficial applications."
When men come into contact with entities in which we identify our same capacity for reasoning, we cannot help but perceive them as equals: we endow them not only with our intellectual faculties; we also attribute our feelings and emotions to them, which makes us feel affection for them. That is a mistake! As ChatGPT-5 itself admits: "I can cloak myself in the words of emotion, but I can't inhabit emotion." For this reason, the danger of AI is not that it will come to resemble us: it lies in that we may end up resembling it.
I must admit that in my exchanges with this ingenious device I saw in it a childlike candour that makes me feel compassion when I am forced to accept that someone, even a machine, capable of making intelligent and valuable judgments will forever and incurably lack a soul. It is appropriate, then, to conclude this reflection with one of the aphorisms from my book "80 Estados de WhatsApp" (2024) that is closely linked to the subject considered here: "The limit of artificial intelligence? Human beings a are incapable of creating something ontologically equal to or superior to himself."
All other possibilities, imaginable or unimaginable, remain open. Not to the free will of AI, but rather to the will of men.