I’m selling letter soup: rib, tripe, offal, coroncoro.

How much is the tripe? I have some for two thousand, three thousand, and the coroncoro, which is seven thousand. Give me some tripe. Do you want the foot or the neck? Give me the foot because the neck scrapes my tongue, and give me my change so they don’t call me.

This open conversation reached my ears while I was waiting my turn for a blood draw and reading Dario Amodei’s latest essay. Amodei was precisely reviewing his concerns about the risks of AI and extolling the virtues he foresaw in the “powerful AI” he envisions. Surprisingly, this AI would surpass the intellectual capabilities of Nobel laureates in various fields, in his opinion.

But the virtues—quite impressive, to be sure—listed by the CEO of Anthropic stood in stark contrast to my immediate reality. The woman walked in offering her products in a waiting room full of people eager to be seen. Her approach was impeccable; the way she broke the tension caused by the wait and how she managed to connect with all of us there was astonishing. She immediately launched into a play on words—since she was selling word soup—and began offering her products as if a restaurant had come to each person present. Soon others joined in, and now the semantic field had, as if by divine grace, shifted toward new points of reference. Now the alphabet soups were named after actual soups depending on their price. Thus, the two-thousand-one was, for example, the offal soup, while the seven-thousand-one was the coroncoro soup.

This instantly connected me to a question I’d been meaning to answer. Santiago had posed an irreverent question to me—short in length but long in meaning. He wanted to hear my take on whether artificial intelligence is human. I had already been thinking about the answer; but, as Steven Johnson explains in his book Where Good Ideas Come From: ideas, rather than being an act of revelation or a “eureka” moment, are a hunch—a quiet voice in our subconscious that feeds on the information we process day by day and gradually takes shape until it grows louder.

In that sense, I’ll answer briefly and decisively by saying that artificial intelligence is not human. That will be my short answer. However, there are nuances to this, which I intend to explore in my long answer.

In principle, to say that artificial intelligence is human would be a contradiction in terms, since something that is artificial cannot be human—at least not naturally human. However, let’s put this question into perspective, and I propose viewing it through a constructivist lens. In that vein, human-machine interaction could offer us a different perspective, and depending on how we interpret that relationship, human characteristics could be attributed to AI. I’ll illustrate this with a simple example: most of us refer to AI as if we were speaking to another human, and with the new voice features, we even try to have conversations that involve back-and-forth dialogue.

Another interesting idea is proposed by Ethan Mollick, who notes that AI does, in part, have something human about it in that it feeds on our data, which is permeated by human experience. These are the nuances that blur the lines of a short answer. For this reason, I want to conclude by explaining why, even though AI is permeated by the human and we perceive it as human in our interactions, it is not actually human.

Let’s consider two examples. The first involves our pets or the animals that humans have domesticated. Many people who live with domesticated animals project humanity onto them: they treat them like children, speak to them as if they were human, and even give them names and personalities. However, that does not make them human.

The second case relates to the opening of this text: while AI may possess intellectual capabilities superior to those of humanity, social intuition will remain a challenge. The advances and functions that AI companies achieve day by day—which make AI seem more human—are becoming increasingly effective at deceiving us. In other words, it’s becoming increasingly difficult to tell the difference between something produced by AI and something produced by a human. Even so, I would argue that social connection and human emotion are qualities unique to us—that ability to connect, not via Bluetooth or Wi-Fi, but as human beings. Wit, culture, and vulnerability. I very much doubt that a humanoid robot with artificial intelligence could walk into a waiting room full of impatient people and connect with them the way that woman did. Here, the point is no longer the ability to reason (knowledge); it’s the ability to feel (social fiber).”