Algorithms: The Ethics That Teaches Machines to Serve Humans

In the accelerated heart of the digital age, where algorithms learn at exponential speed and Artificial Intelligence insinuates itself into every corner of social life, a voice is raised that speaks not the language of profit, but that of dignity. It is the voice of Paolo Benanti , a Franciscan of the Third Order Regular, theologian, and professor of Algorithms at the Pontifical Gregorian University, recently appointed by the UN to the Advisory Council for Global AI Governance.

Paolo Benanti (born July 20, 1973 in Rome) is an Italian priest, theologian, and philosopher of the Third Order Regular of Saint Francis . He teaches at LUISS Guido Carli University and Seattle University and has served as an advisor to Pope Francis on artificial intelligence and the ethics of technology.

Algorithms and the Entropy of Information: Teaching Machines the Weight of Meaning

In his speech at TEDxTreviso, Paolo Benanti posed a radical question: who will teach ethics to machines? In a world where Artificial Intelligence processes data at exponential speed, the risk is not only technical, but semantic: that information loses its human significance, its connection to truth, justice, and dignity.

This is where Shannon entropy comes into play . Introduced in 1948, it is a measure of the uncertainty associated with a random variable. The higher the entropy, the more unpredictable the information. But this unpredictability, in algorithmic logic, is just noise . For humans, however, it is often where meaning hides.

Paolo Benanti invites us to reflect on this gap: the algorithm seeks order, predictability, optimization. But ethics arises precisely where information is ambiguous, where data is not enough. Algorithms , then, are not just a moral grammar for machines: they are an attempt to teach them to recognize that not everything that is information is automatically meaningful.

In this sense, entropy becomes a powerful metaphor. AI reduces uncertainty, but humans live in uncertainty. AI seeks patterns, but humans seek meaning. And meaning, as Benanti reminds us, cannot be calculated: it is contemplated, discerned, and preserved.

Algoretics: A Universal Language for Algorithmic Justice

Benanti doesn't propose a simple ethics applied to technology. He proposes an algor-ethics : a moral grammar capable of translating the value of individuals into code understandable by machines. In a world where algorithms decide who receives a loan, who gets hired, who is monitored, algorithmics is an act of cultural resistance. It's not about curbing innovation, but about directing it. It's about remembering that AI is not an oracle, but a tool. And that every tool, no matter how sophisticated, must remain at the service of humanity.

Algocracy and the Risk of Invisible Power

Paolo Benanti also coined the term algocracy to describe the growing power of algorithms in governing our lives. A power that manifests itself not with uniforms or decrees, but with notifications, recommendations, and scores. It is the power of the invisible, of calculation that decides without explanation. Algoretics opposes this opacity, demanding transparency, accountability, and above all an anthropological vision: the algorithm must understand the human not only as data, but as mystery.

The UN and the global challenge

In 2023, the United Nations recognized the urgent need for global ethical guidance on AI, establishing a committee of 39 experts. Among them, the only Italian is Paolo Benanti . Its task is to contribute to the drafting of universal guidelines that ensure AI does not become a tool of exclusion, but of justice. A task that combines theology, philosophy, technology, and diplomacy.

A spiritual and cultural challenge

The question Paolo poses is radical: will our understanding of what is right and true be able to influence the way algorithms think? It's a question that concerns not only engineers, but every citizen, every educator, every artist. Because AI isn't just a technical issue: it's a spiritual issue. It reflects what we're willing to teach our digital creatures.