Over 200 Nobel Laureates and OpenAI, Anthropic, Google DeepMind Chiefs Debate AI Risk at the Vatican

From July 14 to 16, the papal gardens of Castel Gandolfo outside Rome are hosting a meeting without precedent in the history of AI debates: more than two hundred Nobel laureates, former heads of state and government, and the chiefs of Anthropic, OpenAI and Google DeepMind are sitting at the same table to discuss whether the development of AI could lead humanity toward a catastrophe comparable to Hiroshima.

A Meeting Shadowed by a Papal Encyclical

The assembly is not just another industry conference. Organizers explicitly invoke Pope Leo XIV's first encyclical, issued in May, in which the pope demands that artificial intelligence be subjected to a principle of 'disarmament' - stripped of the capacity to wage war and harm people. The document, titled Magnifica humanitas, addresses the protection of the human person in the age of artificial intelligence.

Vatican adviser Ron Ivey invoked the pope's words at the meeting, according to which AI must remain a tool serving humanity rather than becoming a weapon. That line has become the unofficial motto of the three-day assembly and appears in nearly every report from Castel Gandolfo.

AI must be a tool in human hands, it must serve humanity. It cannot be a weapon - Ron Ivey, Vatican adviser, quoting Pope Leo XIV

Why the Hiroshima Comparison Comes Up

The event's formal name, the World Assembly of Nobel Laureates on Artificial Intelligence and Nuclear War, signals the direction of the discussion: participants are comparing the current moment in AI development to the moment humanity first acquired weapons capable of mass destruction. Underlying the debate is the question of whether AI systems capable of acting autonomously in armed conflicts require a control regime similar to the one imposed on nuclear weapons after 1945.

The question is not merely symbolic. Reports are increasingly surfacing about the use of AI systems in military operation planning and work on autonomous weapons systems, while talks at the United Nations on restricting such weapons have failed to reach agreement for months. The Castel Gandolfo assembly is meant to push that debate to a higher level by directly involving the chiefs of the companies building the most powerful models.

What the Encyclical Says About Power

In Magnifica humanitas, Leo XIV states that technology in itself is neither good nor bad, since it is made by humans, but it is not neutral either, because it takes on the character of those who design, fund, regulate and use it. The pope also writes that truth is a common good, not the property of those who hold power, a line that assembly participants read as a direct critique of the concentration of AI capabilities in the hands of a few tech corporations.

Truth is a common good, not the property of those who hold power - Pope Leo XIV, encyclical Magnifica humanitas

The encyclical also addresses the labor of people mining the rare earth metals needed to build AI infrastructure, calling their situation a new form of enslavement and an ethical challenge of the digital transformation era. That thread ties the AI safety debate to the realities of global hardware supply chains, which are usually discussed separately from ethics.

The Program and the Capitoline Finale

The three days of proceedings in the papal gardens are set to conclude with participants moving to the Capitoline Hill, where on July 16 Cardinal Baldassare Reina will deliver an address followed by the signing of the Rome Declaration for a disarmed and disarming peace in the age of artificial intelligence. The document is meant to represent a joint position of the scientific, political and religious communities on the risks of AI development, though it remains unclear whether it will carry any binding weight for tech companies or governments.

The presence of the Anthropic, OpenAI and Google DeepMind chiefs alongside peace laureates such as Muhammad Yunus and Maria Ressa is meant to show that the makers of the most powerful AI models are willing to discuss limits on their industry in a forum they do not control. Critics of similar initiatives note, however, that earlier declarations and codes of good practice in the AI industry have rarely translated into measurable changes in the pace of deploying new, potentially risky systems.

What It Means Beyond the Vatican

What matters most is that the Castel Gandolfo assembly joins a growing list of international initiatives from 2026 attempting to give AI development a safety framework comparable to the one that has governed nuclear weapons for decades. Similar notes have already been struck earlier this year at the UN summit in Geneva and in statements from British diplomacy, but the Vatican assembly is the first to bring religious, scientific and business authority together in one place at one time.

The effects of the declaration signed on the Capitoline Hill won't be known right away. The key question is whether companies like Anthropic, OpenAI and Google DeepMind, which sent representatives to Castel Gandolfo, will treat the document as more than a diplomatic gesture toward the Vatican and the Nobel laureate community.

Sources: Vatican News (vaticannews.va), Misyjne.pl (misyjne.pl), Dorzeczy (dorzeczy.pl), Interia (wydarzenia.interia.pl)