Artificial intelligence has entered our lives with a speed difficult to compare with previous transformations. Certainly, it is not the first great revolution humanity has ever experienced, but it could be the fastest. Indeed, just a few years ago, many of the tools we now use daily still seemed experimental. Now they write texts, answer questions, organize information, generate images, translate conversations, and accompany the work processes of millions of people. The feeling of accelerated change is real. And we are probably just seeing the beginning.
Following the discussion “Artificial Intelligence and Humanism” organized by the TSAWA Foundation in Madrid on May 7th, 2026, many questions were left open that deserve further reflection beyond the format of the debate. Perhaps the main one is this: what kind of relationship do we want to establish with these technologies?
Often public discussion about AI oscillates between two unhelpful extremes. On one hand, the apocalyptic vision according to which machines will end up replacing us and completely dehumanizing society. On the other, an almost automatic confidence that any technological advance necessarily represents progress. Neither position seems sufficient to really understand the current events. The issue is probably not whether artificial intelligence is good or bad in itself. The issue is what use we want to make of it and what parts of our lives we decide to keep under human responsibility.
When AI is used as a tool for amplification, its potential is enormous. It helps us process more information, work more quickly, access complex knowledge, or develop tasks that previously required much more time and resources. In that sense, it can become a profoundly useful technology. Moreover, it is a type of technology that, at least in large part, could be made accessible to everyone, thus contributing to equal opportunities.
But there is an important difference between amplifying capacities and replacing them. Technology starts to become problematic when we stop using it as collaboration and begin turning it into total delegation. When we withdraw from the scene and let AI think for us, decide for us, or interpret the world in our place, we are accepting avoidable risks that might be very dangerous. Not because machines have bad intentions, but because we renounce exercising certain capacities that until now had characterized us as humans.
This, to some extent, has happened before. There are skills that we have practically lost by neglecting their use. Spatial orientation has weakened since we depend on GPS. Handwriting occupies less and less space in our daily lives. Many simple mental operations have disappeared because we have externalized them completely onto devices. With generative AI, we are approaching something similar, but in much more delicate territories: attention, critical thinking, the independent elaboration of ideas, even certain forms of moral sensitivity.
Machines and their software, at least as we conceive them until now, are a product of human design and creation. They have no intentionality, no free will, no moral compass. It is a human responsibility to give direction to technology, especially regarding morality.
But thinking requires effort, and making decisions perhaps requires even more. And precisely because of that, those capacities strengthen when exercised. If everything that is complex, slow, or ambiguous automatically passes into the hands of artificial systems, there is a risk of losing not only technical skills, but also human depth.
There is another particularly interesting phenomenon: many people already maintain constant conversations with language models as if they were speaking with someone real. And, in a sense, it is understandable. These tools have reached a level of fluency sufficiently sophisticated to generate an illusion of listening, understanding, and companionship. That opens valuable possibilities. Some people use these systems to organize their thoughts, reflect, clarify emotions, or even as therapeutic support. The machine then functions as a kind of structured mirror: it returns our own ideas organized, reformulated, and expanded. But new questions also arise: What happens when we start substituting human bonds for simulated relationships? What does it imply to confide intimacies to systems designed by large technology corporations? To what extent does the feeling of understanding that these tools produce correspond to any actual understanding?
Perhaps one of the keys lies in remembering something very simple: machines can process language, but they do not live. They have no human experience, vulnerability, consciousness, or compassion. They do not feel joy, or sorrow, or fear, or satisfaction. They can imitate many external expressions of our qualities and feelings, but that does not mean they possess any of this in a genuine sense. That is why it is absolutely crucial that we do not disappear in our relationship with AI.
Artificial intelligence can be an extraordinary tool as long as there remains a human intention behind it. What is truly decisive is not that machines are becoming increasingly intelligent, but that we do not stop being conscious, critical, and fully human. Perhaps the challenge of this era consists not only in developing more advanced technologies, but also in developing a culture capable of using them without losing precisely that which machines cannot replace.