Imagine you are in an autonomous car. Suddenly, an unavoidable situation arises: a pedestrian crosses out of place and, at the same time, another car invades your lane. The car must decide what to do. It is not a simple question of braking or accelerating, but of facing a moral dilemma. How should an agent act in such a case?

So far we have seen agents that react, plan, learn and even cooperate with each other. But the future points to something deeper: cognitive agents, capable of reasoning at levels closer to the human mind. These agents not only handle information and goals, but build models of themselves, simulate emotions and can understand the perspective of others.

The key is to endow them with capabilities that we now associate with human cognition:

• Self model: the agent understands his own limitations and strengths, and can represent himself within the environment.
• Theory of mind: the ability to infer what other agents (or people) know, feel or want.
• Causal reasoning: not only to detect correlations, but to understand causes and consequences.
• Ethical judgment: making decisions considering values and possible moral dilemmas.

These advances open the door to agents that not only optimize tasks, but also reflect on how and why they make certain decisions. For example, a healthcare system could recommend a treatment not only based on medical data, but also taking into account the patient’s social and emotional situation. Or an educational assistant might adapt the tone of his or her responses based on perceived motivation or frustration in the learner.

At SMS Sudamérica we believe that cognitive agents will be a turning point. We will move from systems that automate tasks to true digital collaborators capable of holding more human conversations, anticipating needs and supporting critical decisions. The challenge will be enormous: establishing clear ethical frameworks, designing oversight mechanisms and ensuring that agent autonomy is always aligned with human well-being.

Because if the dilemmas of the future show us anything, it is that intelligence alone will not be enough. We will need agents who combine calculation with simulated empathy, strategy with responsibility, autonomy with transparency.

In short, cognitive agents represent the most ambitious frontier of this technological revolution: intelligences that not only solve problems, but also understand their context, their consequences and their impact on the world we share.

With this episode we close the second season of Historias del Futuro, Hoy. But more than an ending, it is a starting point: the story of the agents is just beginning, and the future invites us to write it together.

Note by: María Dovale Pérez