Pope Leo XIV’s first encyclical, Magnifica Humanitas, makes artificial intelligence a matter of human dignity, public policy and democratic control. The document argues that AI should serve human flourishing rather than quietly persuading people to treat human limits as defects and machine intelligence as a replacement for human judgement.
Watch Desk analysis
What happened
The Observer’s account of the encyclical says the Pope identifies three immediate dangers from AI: the ease of obtaining results, which may weaken creativity and judgement; the apparent objectivity of machine responses, which can hide blind spots; and systems’ imitation of human connection, which may erode genuine relationships.
The document does not reject the technology. Instead, it calls for rules that make automated decision-making more transparent, protect personal data and prevent AI’s future being left solely to unelected technology leaders. It also takes aim at transhumanist ideas that imagine humanity being uploaded, upgraded or made redundant by synthetic intelligence.
That is a notably broad intervention. This is not merely a plea for safer software or better chatbot manners. It is an argument about who gets to define progress, whose interests technology serves and whether societies should accept limits as part of being people rather than problems awaiting a firmware update.
Why it matters
AI policy is often discussed in the language of productivity, national competition and model capability. The Pope’s intervention shifts the emphasis towards dependence, relationships, vulnerability and power. Those are not technical specifications, but they are precisely the conditions in which technical systems operate.
The practical concern is easy to recognise. A system that produces an answer quickly may save time, but it can also discourage someone from thinking through a problem. A system that sounds confident may be trusted in hiring, welfare, healthcare or warfare even when its reasoning is opaque or wrong. A system that mimics warmth may offer companionship while leaving people less practised at building relationships with one another.
The encyclical is also a reminder that AI governance will not be written only by engineers, regulators and investors. Faith leaders, artists, trade unions, teachers and ordinary users all have a stake in deciding what should remain human-led. The technology industry may prefer a narrower conversation. It is unlikely to get one.
Our read
The useful part of this intervention is not the religious packaging, which will persuade some readers and irritate others. It is the insistence that efficiency is not the same as wellbeing, and that a technically impressive system can still be socially impoverishing.
The Pope is right to demand transparency and public rules. But moral language must eventually meet measurable safeguards. If a government says an AI decision is explainable, people need to know what explanation they receive, who can challenge it and what happens when the system is wrong. If a company says its product strengthens human agency, it should show how that claim is tested rather than placing it on a tasteful background image.
The encyclical’s larger challenge is whether societies can value human judgement without romanticising every human institution. People are capable of bias, cruelty and spectacular administrative incompetence too. The answer cannot be to worship either people or machines. It should be to decide where machines help, where they must be constrained and where responsibility cannot be delegated at all.
What to watch
- Policy uptake
Whether lawmakers borrow the encyclical’s language on transparency, data protection and human control. - Industry response
Whether AI companies answer its concerns with testable safeguards rather than broad assurances. - Human connection
How schools, workplaces and care services measure the effects of AI-mediated interaction. - Public debate
Whether the conversation expands beyond capability and investment to include dignity, dependency and power. The argument is now larger than whether AI can do useful things. It is whether usefulness is enough to justify handing it more of the decisions and relationships that make a life recognisably ours.
Discussion spark: Should AI policy begin with measurable human outcomes such as autonomy, judgement and relationships, or is that too subjective to constrain rapidly changing technology?
Sources and evidence
- The Pope vs the cult of AI – observer.co.uk (16 September 2026, 08:53 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.