The Most Important AI Ethics Document of 2026 Was Not Written by a Tech Company

The Vatican published a moral governance framework for the age of AI today — and the business leaders it is most directly addressed to are the least likely to read it.

Sign up for my Substack daily AI newsletter here.

See my AI Training course portfolio for corporate Business Leaders here.

Article link: https://open.substack.com/pub/johanosteyn/p/the-most-important-ai-ethics-document?r=73gqa&utm_campaign=post&utm_medium=web&showWelcomeOnShare=true

Follow me on LinkedIn: https://www.linkedin.com/in/johanosteyn/

On 15 May 2026, Pope Leo XIV signed a document. He dated it deliberately. That day was the 135th anniversary of the publication of Rerum Novarum, the landmark 1891 encyclical in which his namesake, Pope Leo XIII, responded to the upheaval of the Industrial Revolution with a moral argument that eventually became the foundation of modern labour law. The parallel was not incidental. It was a statement of intent.

The document, titled Magnifica Humanitas, was published and formally presented at the Vatican’s Synod Hall in the most unusual governance event of the year: a Pope, two cardinals, two theologians, and a co-founder of one of the world’s leading AI safety companies, standing together to present what may be the most substantive moral framework for the age of artificial intelligence that any institution has yet produced.

The business leaders who most need to engage with it are the least likely to open it. That is their most expensive mistake.

CONTEXT AND BACKGROUND

To understand what Magnifica Humanitas is and why it matters beyond its religious context, it is necessary to understand what Rerum Novarum was and what it eventually produced. In 1891, industrialisation had already reshaped work, concentrated wealth in the hands of those who owned the machines, displaced workers without adequate social protection, and generated the structural conditions for social instability that governments and corporations had largely failed to address. Pope Leo XIII’s encyclical condemned laissez-faire capitalism without endorsing socialism, articulated the rights of workers to form unions and bargain collectively, established the principle of a living wage, and placed human dignity at the centre of the moral framework for how economic transformation should be governed. It was not immediately translated into law. It entered the moral argument, shaped the social movements that demanded legislation, and became the foundational principles of the labour law frameworks that govern employment across the world today.

Pope Leo XIV chose his name at his election in May 2025 to signal continuity with that tradition. In his first public statement as Pope, he drew an explicit connection between Rerum Novarum’s teaching on labour and the rise of artificial intelligence, describing AI as posing new challenges for the defence of human dignity, justice, and labour. Magnifica Humanitas is the fulfilment of that signal. It centres on the protection of the human person in the age of AI, the obligations of those who build and deploy the technology, the distribution of AI’s benefits and costs, and the moral framework within which decisions about AI should be made.

The presence of Christopher Olah, co-founder of Anthropic and lead of its interpretability team, at the formal presentation is the detail that most deserves examination. Anthropic is not a neutral choice. Earlier in 2026, the company refused to remove safeguards preventing its AI models from being used for lethal autonomous warfare without human oversight. The Pentagon designated it a supply chain risk as a result — the first time that designation was applied to a US technology company.

OpenAI signed the defence contract that Anthropic declined. The Vatican’s decision to present its moral framework for AI governance alongside a co-founder of the company that held a principled line at considerable commercial cost is not a public relations exercise. It is a signal about where the moral centre of gravity in AI governance is located.

INSIGHT AND ANALYSIS

The AI governance landscape of 2026 is not short of documents. The EU AI Act classifies risks. The NIST AI Risk Management Framework assesses trustworthiness. National AI strategies articulate ambitions. Corporate AI ethics policies describe principles. What almost none of them do is ask the distributional question: who captures the gains of AI productivity, and who bears the costs of AI displacement? That question has been consistently deferred in every major governance conversation of the past five years, because the answer is politically and commercially inconvenient for the institutions with the most power to provide it.

Magnifica Humanitas asks it directly. The encyclical situates AI within the same moral tradition that asked the same question about industrial capitalism in 1891, and it draws on 135 years of developed social teaching, the principle of human dignity as inviolable, the priority of labour over capital, the universal destination of the goods of the earth, the obligation of solidarity with those who bear the costs of transformation, to construct an answer that is both more ambitious and more precise than anything the technology industry has produced in its own ethics work.

I have previously written about the distributional question as it manifests in employment law, examining the Chinese courts’ ruling that a company’s strategic choice to adopt AI cannot simply transfer the costs of that choice to the workers whose roles it eliminates. Magnifica Humanitas places that legal argument in its moral context. The obligation to ensure that the gains of AI transformation are shared and the costs are not simply transferred to those with the least power to resist them is not merely a legal requirement in some jurisdictions. It is a moral obligation that precedes and outlasts any specific regulatory framework.

The historical parallel that the encyclical’s timing invites is not rhetorical. The industrialists who ignored Rerum Novarum in 1891 did not avoid its consequences. They deferred them. The social movements it helped shape, combined with the political pressures those movements generated, produced the labour legislation that eventually governed their industries in ways they had not anticipated and could not reverse. The organisations that engaged with the moral argument early were better positioned for what followed. Those who dismissed it as ecclesiastical commentary were not.

IMPLICATIONS

For boards and executives, the strategic question Magnifica Humanitas raises is not theological. It is governance. The encyclical asks whether the people whose working lives are being reshaped by your AI deployment have been treated with dignity. Whether the productivity gains your organisation is capturing are being shared in ways that reflect the contribution of the workers whose roles made those gains possible. Whether the decisions being made about AI deployment in your organisation are being made with genuine accountability to the people they affect, or whether they are being made with accountability only to shareholders and senior leadership.

These are not questions that any existing AI governance framework requires you to answer. They are questions that the moral argument now circulating at the highest levels of global discourse, from the Vatican to the Chinese courts to the European Parliament, is beginning to require organisations to engage with. The trajectory from moral argument to social pressure to regulatory obligation is the one Rerum Novarum traced over the course of the twentieth century. Magnifica Humanitas is at the beginning of that trajectory today.

US Vice President JD Vance stated publicly that he was looking forward to reading the encyclical. That a senior figure in the administration most aggressively pursuing AI deregulation is treating a papal document on AI as worth engaging with is itself a signal. The encyclical has reached an audience well beyond Catholic institutions. It is being taken seriously across ideological and institutional lines by people who will shape the regulatory environment in which your organisation’s AI decisions will eventually be judged. The question for business leaders is whether they can afford to be less informed about the most significant moral governance framework produced for AI in 2026 than the politicians who will eventually regulate them.

CLOSING TAKEAWAY

Five years of corporate AI ethics work has produced principles without accountability, frameworks without enforcement, and documents that describe aspirations rather than obligations. The institution that has spent 135 years developing a moral response to what transformative technology owes to the people it displaces published its answer today. It does not classify your AI model by risk tier. It does not ask whether your training data is clean. It asks whether the people whose lives your AI shapes have been treated as ends in themselves or as means to someone else’s productivity gain. That question will not remain in the domain of moral philosophy indefinitely. It will follow the path that Rerum Novarum traced. The organisations that engage with it now, before it becomes law, will be better positioned than those that wait for it to arrive in a regulatory instrument they can no longer ignore.

Johan Steyn is a prominent AI thought leader, speaker, and author with a deep understanding of artificial intelligence’s impact on business and society. He is passionate about ethical AI development and its role in shaping a better future. Find out more about Johan’s work at https://www.aiforbusiness.net