And if AI agrees with us too much?

BarcelonaYou have just argued with your partner and are convinced that you are right. Instead of explaining it to a friend, you open an artificial intelligence application and describe everything that happened to it. The answer does not wait: it understands how you feel, validates your discomfort, and explains to you why your reaction is so understandable. It might even suggest a message for you to send to them explaining how you feel. When you close the device, it is easy to leave with the feeling that someone has truly listened to you and that, moreover, you were the one who was completely right.

But what if the problem is precisely that the machine has not disagreed with you? “I believe that what is good about a psychologist is that they make you see uncomfortable situations and make you think,” says Ariadna Vilalta, cyberpsychologist and author of A Life Always Online (Destino, 2026). Her concern is not that artificial intelligence listens to us or helps us, but what might happen if we get used to interacting with technology that offers us immediate answers, that does not get impatient, and that, above all, often ends up telling us what we want to hear.

AI is offering us precisely the opposite of what happens in human relationships, which force us to negotiate, wait, interpret silences, and, sometimes, accept that the other person does not think like us or that they are right. Nevertheless, Vilalta believes that there is a question here that we still do not know how to answer: what will happen to our tolerance for discomfort if we increasingly have more tools capable of eliminating it immediately?

In fact, it is not a question exclusive to artificial intelligence, states Vilalta. Internet has been getting us used to living in immediacy for years now. It informs us quickly when we want to know something, distracts us when we are bored, connects us when we feel lonely, and allows us to respond to any message in a matter of seconds.

“We have been living with the internet for many years, and the problem is that we have done so without having been educated on it,” she points out. While cyberpsychology—the area that studies how technology affects our behavior and our mental health—is still a relatively new field, technology has long since fully entered into our way of relating, thinking, and working. “We are late,” assures Vilalta.

The danger of stopping choosing

Another of Vilalta's concerns has to do with something apparently more innocent: recommendations. “Before, you would go to the cinema and look at the listings to read the synopses and, based on that, you would evaluate what you would watch,” she explains. Now, however, a platform tells us directly what we might like. It doesn't seem like a big difference, but if we systematically let algorithms decide what we read, what we watch, or what we listen to, we can also end up watching more or less the same thing every time. “We stop choosing for ourselves a little bit,” she summarizes.

At this point, we already know that algorithms are designed to predict what will keep us in front of the screen longer and to offer us content that we will likely be interested in. The result can be very comfortable, but it can also reduce the possibility of encountering what we were not looking for. And this, according to Vilalta, directly affects critical thinking. “I don't believe you can form a solid opinion if you haven't seen A, B, and C and decided what is best for you,” points out the cyberpsychologist.

Faced with all this, Vilalta advocates for recovering things that may seem untechnological: reading, writing, studying philosophy and ethics, discussing ideas, and learning to argue. “We lack that point of being critical, like when we used to have debates in school, of giving everything a second thought,” she states. And she considers this training especially important now that we have tools capable of generating answers on practically any topic. Because thinking does not consist only of obtaining an answer. It also implies doubting, contrasting information, changing one's mind, and enduring the fact of not knowing what to think for a while.

This is, perhaps, one of the paradoxes of artificial intelligence, explains Vilalta. It can help us access more information than ever, but that does not necessarily mean that we have more discernment. It can explain anything to us in seconds, but it cannot decide for us what is relevant, what is true, or what we want to think.

“With AI we are heading for the worse,” says Vilalta when talking about the loss of critical thinking. Not because she considers technology to be bad. In fact, she insists that it is not about rejecting it: technology has enormous potential and she herself works with it. The problem, she says, is that we hadn't even learned to live with the internet when the next revolution, that of AI, had already arrived.

And this time perhaps the question is not just how much time we spend in front of a screen, but how much of our time, of our decisions and of our thinking we are willing to delegate to it.