Vatican Releases First Encyclical on AI with Anthropic Co-founder on Stage

Pope Leo XIV published the encyclical Magnifica Humanitas on Monday, 25 May, the first magisterial document from the Catholic Church dedicated solely to artificial intelligence. The presentation took place at 11:30 in the Synod Hall at the Vatican, with the personal presence of the pontiff, an atypical occurrence that analysts from the Holy See interpreted as a direct priority signal from Leo XIV. Christopher Olah, co-founder of Anthropic and director of research in model interpretability, was the only AI industry executive on the panel. Cardinals Víctor Manuel Fernández and Michael Czerny, along with theologians Anna Rowlands from Durham University, and Léocadie Lushombo from the Jesuit School of Theology in Santa Clara, made up the rest of the table.

Olah's Selection and the Signal Sent to the Sector

By inviting Olah as the sole representative from an AI company, the Vatican specifically signalled that interpretability research, a discipline focused on understanding the internal mechanisms of large-scale language models, is the technical approach the Church deems compatible with the protection of human dignity. Olah co-founded Anthropic in 2021 and has devoted his research career to making AI models understandable to humans, in contrast to security strategies that rely solely on behavioural alignment. Anthropic operates the Claude Mythos Preview through Project Glasswing, a cybersecurity initiative with around 50 partner organisations supported by up to $100 million in usage credits provided by the company itself. In its first month, the project identified over 10,000 critical vulnerabilities in widely used infrastructure systems, including operating systems and browsers. The company formed a global alliance with KPMG in 2026 to embed Claude models into the Digital Gateway, the firm's consulting services platform. That the Vatican, with 1.4 billion faithful, chose Olah for the stage conveys a technical preference, not just a diplomatic invitation.

The Parallel with Rerum Novarum

Leo XIV signed Magnifica Humanitas on 15 May, exactly 135 years after Leo XIII signed Rerum Novarum, the encyclical that defined the Church’s social doctrine in response to the Industrial Revolution. The Rerum Novarum of 1891 became the foundation for labour legislation in dozens of countries throughout the 20th century and continues to be cited in international treaties and public policy debates. For the 1.4 billion Catholics worldwide, Rerum Novarum was a moral instruction that, over time, led to concrete political pressure on states and employers. The current pope frames AI as a disruption of comparable magnitude to industrialisation. For consulting firms advising clients on automation initiatives, this framing has concrete implications: in any market where Catholic social doctrine guides corporate decision-making, including parts of Western Europe, Latin America, and the Philippines, Magnifica Humanitas will serve as a reference in discussions about AI governance in the upcoming decades.

Concrete Positions: Work, Platforms, and Armed Conflict

The encyclical does not condemn AI. According to sources with access to the text prior to its publication, the document starts from the premise that technology is morally neutral but conditions its legitimacy on the equitable distribution of the productivity gains it generates. Operationally, Magnifica Humanitas advocates for the workers' right to retraining and economic transition support when replaced by automation, as well as proportional participation in the profits generated by systems that have taken over their roles. The document specifically criticises recommendation systems designed to maximise usage time at the expense of user well-being, describing this design choice as moral damage, not just a product decision. It also calls for international regulation of AI use in military operations.

What Executives and Consulting Firms Need to Monitor

Encyclicals do not create legal obligations. The relevance for compliance teams is indirect but present. European regulatory experts have indicated that the text will be cited in parliamentary debates concerning provisions of the EU AI Act related to high-risk systems impacting workers and the military use of AI. The European Parliament is expected to formally adopt the AI Omnibus in July, with new rules coming into effect in August 2026. In sectors such as finance and healthcare, the process of ethical due diligence in AI contracts is rapidly expanding. That Anthropic participated actively in the launch while operating partnerships with governmental organisations and firms such as KPMG reinforces the reading that ethical positioning and security in AI platforms will become a competitive differentiator in procurement. Firms like BCG, Deloitte, and Accenture, which have published their own frameworks for AI ethics, will need to position these frameworks in relation to the reference the Vatican has just established for 1.4 billion people.