Léon XIV Calls for Responsible AI in Business

The Encyclical of Leo XIV: A Call for Responsibility in the Face of AI

In his encyclical Magnifica Humanitas, Leo XIV addresses the crucial issue of protecting human dignity in the age of artificial intelligence. This text comes at a time when AI is moving beyond laboratories and IT departments to fully integrate into business processes, human resources, finance, marketing, and even managerial decision-making.

Leo XIV emphasizes that technology, while not inherently harmful, can become an instrument of domination if not directed toward the common good. This reflection is aimed at all leaders, urging them to ask not whether to use AI, but how it can make the enterprise more humane, just, and intelligent.

Automate, Control, Humanize: A Governance Method

In his work on AI applied to management, one approach is often summarized by the formula: automate what is ordinary, control what is extraordinary, humanize what is essential. This method is not merely a moral stance but a genuine governance strategy. It helps avoid two pitfalls: technological fear that paralyzes and technological idolization that too quickly delegates to machines what still requires human discernment.

AI is Not Neutral: It Embodies a Worldview

One of the major contributions of Magnifica Humanitas is the reminder that AI cannot be considered morally neutral. Leo XIV points out that every technical device involves choices: what it measures, what it ignores, what it optimizes, and how it classifies people and situations.

For managers, this idea should become a reflex in decision-making. For example, a commercial scoring algorithm does not measure a customer's "value," but rather what the company has decided to call value. Similarly, an HR tool does not measure a candidate's "potential," but classifies certain signals according to defined, weighted, and trained criteria.

AI in Judgment Chains: A Latent Danger

Companies must stop referring to AI as a mere tool. Unlike a hammer or a screwdriver, AI enters judgment chains, influencing, recommending, prioritizing, and sometimes excluding. It does not always formally decide, but it can steer decisions to the point of making them almost automatic.

The danger is not that AI will abruptly replace the leader, but that it will gradually instill a laziness in judgment. The manager no longer says "I decide," but rather "the tool recommends." The recruiter no longer states "I take responsibility for this choice," but says "the score is low." The salesperson no longer claims "I know this customer," but states "the algorithm classifies them as low priority."

Augmented Management vs. Delegated Management

Leo XIV stresses that AI systems lack moral consciousness, lived experience, and the capacity to bear the weight of consequences. They can imitate language, simulate empathy, produce evaluations, but they do not understand what they produce in the human sense.

This distinction is fundamental for businesses, drawing the line between augmented management and delegated management. Augmented management uses AI to see better, compare better, detect weak signals better, and structure information better. It saves time on the ordinary to devote more energy to what requires judgment: arbitrating, listening, encouraging, correcting, negotiating, deciding.

Delegated management, on the other hand, transfers an increasing share of responsibility to AI without a clear framework. It automates because it is possible, not because it is right. It confuses apparent precision with truth. It transforms algorithmic recommendations into verdicts. It makes the tool a silent hierarchical superior.

Traceability and Explainability of Decisions

In a responsible company, every sensitive decision must remain traceable, explainable, and contestable. Who validated it? On what data? With what limits? What recourse is available? What human oversight is there? What are the consequences for the person involved?

Leo XIV applies this requirement to decisions affecting work, credit, access to services, reputation, or individual freedom. In the business context, this directly concerns recruitment, performance evaluation, commercial prioritization, financial analysis, customer service, public communication, risk detection, and talent management.

The Risk of Accelerated Production Without Discernment

Since the massive arrival of generative AI, many companies have approached the topic through tools: which chatbot to use? Which image generator to test? Which assistant to integrate? What time savings to expect? These questions are useful but insufficient.

The strategic question is deeper: what do we do with the time saved?

If AI is used solely to produce more content, more emails, more reporting, more sales solicitations, more noise, then it risks accelerating mediocrity. It will make the company faster, but not necessarily smarter. It will increase volume, but not necessarily value.

Conversely, if AI allows teams to be freed from repetitive tasks to enhance analysis, creativity, customer relations, service quality, and managerial discernment, then it becomes a lever for transformation.

The Pope writes that the speed and ease with which AI provides answers can weaken personal judgment and creativity if it leads us to delegate too much. For managers, this is a very concrete warning: AI should not become a machine to avoid intellectual effort. It should become a machine to better direct that effort.

AI Governance: A General Management Challenge

Leo XIV speaks of responsibility, transparency, and AI governance. These three words should now be on the agenda of management committees.

AI governance is not about drafting a charter of good intentions that no one reads. It is about defining operational rules.

First rule: map the uses. Many leaders are already unaware of how many AI tools are being used in their company. Employees test, copy-paste, automate, generate, translate, summarize. This is normal. But without mapping, there is no control.

Second rule: classify the risks. Not all uses are equal. Using AI to rephrase an internal note does not carry the same level of risk as using it to analyze applications, respond to vulnerable clients, produce a financial recommendation, or handle confidential data.

Third rule: define prohibited or strictly regulated areas. Certain decisions must remain human by principle. Others can be assisted but never automated without oversight. Still others can be largely automated because they are repetitive, low-sensitivity, and easily verifiable.

Fourth rule: train the teams. A company that adopts AI without training creates dependency. It provides a powerful tool without teaching how to use it. It increases apparent productivity but undermines actual competence.

Fifth rule: establish a right to doubt. An employee should be able to say: “I do not understand this recommendation,” “I think the model is wrong,” “This answer is plausible but dangerous,” “This data should not be used.”

Governance is not a hindrance to innovation. It is the condition for trust. And without trust, there will be no sustainable adoption.

Europe and Trustworthy AI: A Strategic Asset

In the encyclical, Leo XIV warns about the concentration of digital power. Control over platforms, infrastructures, data, and computing power increasingly belongs to large economic and technological players, often more powerful than some states. This concentration can become opaque and escape public control.

This point is strategic for European companies. We sometimes tend to think that Europe is lagging behind because it regulates more. But regarding AI, the framework can become a competitive advantage.

In a world saturated with powerful tools, trust becomes an economic value. Customers will want to know how their data is used. Transparent and responsible governance systems can thus become a major asset for companies wishing to stand out in an increasingly demanding market.