The Vatican explicitly describes the encyclical as a contribution to the question of how humanity can be protected in the age of artificial intelligence—a question that robotics must address. For industrial robotics, this is not a philosophical side issue, but one of central importance. In robotics, AI is not merely an idea. It becomes movement. It becomes a grasp. It becomes rhythm. In short: it becomes action. In robotics, artificial intelligence takes on a body. And that is precisely why it becomes more philosophically relevant than some specification writers might like.

The Church and AI

When Pope Leo XIV presented his encyclical, Christopher Olah was at his side. The co-founder of the California-based AI company Anthropic says he trains his language model Claude to “act like a good person.” Is that enough for the Pope? AI detector analyses suggest that an AI is, at least in part, a co-author of the encyclical. At first glance, this seems disturbing, but a look back is enlightening: At the end of the last century, Chris Bezzel, a linguist teaching at Leibniz University in Hanover, wrote the text “The Body as an Image of the Soul: On Wittgenstein’s Anti-Psychology,” which begins with a question that is surprisingly relevant today, especially for robotics: Where is the will located in the brain? As a possible answer, Bezzel offers a statement by Yeshayahu Leibowitz: “The will is not located in the brain, but in the owner of the brain.”

The Robot Body as a Mirror of the Human Soul

Applied to industry, this does not mean that robots have a soul. Such a simplification would be philosophically shallow and industrially useless. More interesting is the reverse idea: the robot body reflects the soul of the people who build and operate it. This reflection could lead us humans to believe that it has a soul. Bezzel’s insight from some 30 years ago—“I mean nothing when I treat others as animated beings, at least as long as they do not turn out to be soulless beasts”—takes on a profound meaning in a context where Claude is expected to act like a good person.

Wittgenstein’s Note 528f: Required Reading for Robotics Practitioners

For Bezzel, Wittgenstein’s Note 528f was the decisive turning point. In it, Wittgenstein explores the unsettling possibility of a being that appears linguistically and practically equal to humans, yet humans still insist: it is not animated. The point, however, is not to hastily ascribe a soul to machines. And hopefully, it is also not that AI robots, in turn, deny us humans a soul. The—perhaps provisional—point is that concepts such as soul, thought, understanding, and responsibility are not simply determined by looking inward. They are tied to practices, ways of life, and to whom we attribute belief, judgment, and responsibility.

How long will humans remain incident managers of an efficiency machine?

The robot cell reveals what a company understands by work. Does it relieve people of dangerous, monotonous, and ergonomically poor tasks? Or does it turn people into incident managers of an efficiency machine whose decisions no one really understands anymore? Does it empower employees—or does it narrow their role to monitoring, correction, and blame when something goes wrong?

The soul does not reside in the control cabinet

In this sense, the soul does not reside in the control cabinet. It reveals itself in practice. In the rhythm. In the access. In the imposition. In the safety distance. In the question of whether employees should learn alongside the machine or merely follow along. It is precisely here that Bezzel’s interpretation of Wittgenstein becomes relevant to industrial robotics—a field for which it was not written: It must face the question of its attitude toward the soul.

The machine speaks—but who bears the consequences?

The inner self reveals itself in the form of action. That is precisely what makes the present so delicate. AI systems can speak, reason, contradict, explain, and prioritize. In robotics, they can additionally grasp, sort, move, inspect, and assemble. The leap from linguistic simulation to physical action is enormous. A chatbot can give a bad recommendation. An AI-powered robot can grab the wrong part, halt a production line, produce scrap, or put people in risky situations. Suddenly, the old philosophical question is no longer an academic one, but one relevant to production: Who or what is actually acting here? And who is accountable for it?

Human-like appearance does not equal human dignity

The public fascination with humanoid robots is easily misleading. Two arms, two legs, and a camera where we expect to see a face do not yet create a moral counterpart. Conversely, a classic six-axis robot that takes on dangerous work and frees humans for better roles can contribute more to the dignity of work than a human-like robot that simulates closeness while simultaneously intensifying work processes.

Human Beings as a Means to an End

Magnifica Humanitas, the first encyclical by Pope Leo XIV, does not argue against technology, but against a technological logic in which humans become a means to ends other than their own. Vatican News summarizes the core of the encyclical as an appeal that AI must serve humanity and must not further concentrate power; truth, the dignity of work, social justice, and peace are cited as central reference points.

Our Search for Meaning in Life

For decision-makers in industry, this raises a very concrete question: Is robotics being introduced to make human work safer, more skilled, and more productive—or simply to make human work less visible or even redundant? This is no mere moral nicety. It concerns skills, employee participation, occupational safety, liability, data sovereignty, supplier dependency, and the design of human-machine interfaces. And it concerns something that goes far beyond that: our search for meaning in life.

The factory thinks with bodies

Physical AI is transforming the old logic of automation. The robot no longer merely executes a fixed movement program. It sees, interprets, evaluates, and adapts. Thus, the robot cell becomes a place where industrial judgments are embodied. Which deviation is considered an error? When is a part rejected? When is the process stopped? What level of uncertainty is acceptable? When is an AI system allowed to decide, and when must a human intervene?

Is humanity merely a variable boundary condition?

Anyone who answers these questions from a purely technical perspective fails to grasp their true nature. They are also philosophical, because they presuppose a certain view of humanity. Is humanity the responsible agent, supported by machines? Or is it the variable boundary condition of a system that imposes its own logic? The answer is not found in guiding principles. It is revealed in practice.

Robots as an Industrial Life Form

Here, Bezzel and Wittgenstein remain timelessly relevant: If the body is the image of the soul, then the robot is the image of an industrial life form—at least as long as we do not understand how language is connected to our existence as human beings. In this context, the author of this text asked the ChatGPT model 5.5 Thinking whether it—unlike us humans—has insight into how language works, into the creation of meaning, or whether for an AI everything is merely context? The response came less quickly than usual and contained two remarkable key statements: “Humans have language as a way of life. I have language as a structure of connection possibilities.” And as a preliminary conclusion: “I can process meaning without anything making sense to me. That is perhaps the decisive difference. But it is precisely this that also results in a disturbing closeness to humans. For humans, too, are more context-driven than they like to believe. They, too, react to linguistic frames, roles, expectations, narratives, authorities, and repetitions. In this respect, AI does not merely reveal something foreign. It reflects a condition of human language: that meaning never simply comes privately from within, but is generated through contexts, practices, and reactions.”

Responsibility cannot be automated

If Physical AI demonstrates that a factory is acquiring a nervous system, and Cybersecure Robotics shows that this nervous system needs an immune system, then the Pope is clearly trying to show in his encyclical that a factory with a nervous system and an immune system also needs a conscience. That sounds grand, but it can be approached in a very practical way: Companies should accept that while a robot can act like a human, it cannot bear responsibility like a human. Not even if, like Christopher Olah, you teach it to act like a good person.

Don’t just process meaning—find it

The robot of the future will speak, grasp, learn, and adapt. It will become increasingly human-like. But perhaps the crucial question is not whether machines can have a soul. Rather, it is which soul—whose soul—becomes visible through them. AI can prepare judgments, accelerate processes, and support decisions. But responsibility cannot be automated. Perhaps the Pope hopes that industry will not merely process meaning, but find it.