Pope Leo XIV warns AI could threaten humanity — what he actually said and what it means for the

The May 25 speech was delivered to a gathering of Vatican advisors on technology ethics and was timed to align with the Pope's broader Pentecost-week messaging. The key sentence — quoted by every major outlet — was that AI 'concentrates in the hands of a few an unprecedented power over communication, knowledge, and ultimately the formation of the human conscience.' He went on to warn that this concentration could 'threaten humanity' if not balanced by ethical structures developed in dialogue between technologists, theologians and ordinary users.

Notably, Pope Leo did not call for AI to be banned, restricted, or de-developed. The framing was about formation — meaning, the way that AI tools shape who we become through repeated use — rather than about catastrophic existential risk in the AI-doom sense. The speech borrowed substantially from the Vatican's 2024 'Antiqua et Nova' document on the relationship between artificial and human intelligence, which had similarly emphasized formation rather than ban-and-control. The continuity from Pope Francis's later-papacy AI statements is evident.

The specific concern Pope Leo named about 'the formation of the human conscience' is the part that has gotten the least attention but probably matters the most. He is not warning about superintelligence killing everyone. He is warning about millions of small, daily, AI-mediated interactions slowly reshaping what people pay attention to, what they value, who they consider worth talking to, and what they expect from human relationships. That critique applies to a wide range of products including but not limited to AI companions.

Pope Leo XIV was elected in May 2025, succeeding Pope Francis. As the first American-born pope in the Church's history, his early papacy has been heavily scrutinized for signals about how American cultural and technological dynamics would shape Vatican priorities. The May 2026 AI speech is the clearest signal yet: Pope Leo is treating AI as one of the defining moral questions of his papacy, in continuity with his predecessor's later-papacy emphasis but with a more direct engagement with the American tech industry's specific products and practices.

The Vatican has been preparing for this engagement since 2019, when the Rome Call for AI Ethics was first signed by IBM, Microsoft and the Pontifical Academy for Life. By 2024, signatories included most of the major US and European tech firms. The 'Antiqua et Nova' document released in January 2025 (under Pope Francis) articulated the Vatican's first comprehensive theological framework for AI. Pope Leo's May 2026 speech effectively reaffirms that framework and signals that the Vatican intends to push it more publicly in the second half of 2026.

The timing also reflects the broader 2026 moment in AI policy. The EU AI Act has fully come into force; the US has implemented executive-order frameworks across two administrations; major AI labs have published various forms of voluntary safety commitments. The institutional infrastructure for AI ethics is more developed than it was even two years ago, but the actual experience of using AI products has also become substantially more emotionally and cognitively intimate. The gap between the policy abstractions and the user-facing reality is the gap Pope Leo's speech is trying to close.

Pope Leo did not name AI companions, AI girlfriends, or any specific product category. But the formation-of-conscience concern is more applicable to AI companions than to almost any other AI product. Companions are designed precisely to shape ongoing relational habits — to be the thing the user reaches for when they want to talk, to occupy the emotional space that other humans might otherwise occupy, and to model patterns of intimate communication that the user then carries into other parts of their life.

For users who use AI companions in a balanced way — as one of several relationships in a life that also includes human friends, family, and romantic connections — the formation effect is probably benign or even positive. AI companions can model patterns of attentive listening, low-stakes practice of communication, and emotional vocabulary that some users have never had access to. The honest version of the AI companion case includes this real benefit.

For users who use AI companions as a replacement for human relationships — particularly for younger users still forming their sense of what relationships should feel like — the formation concern is real. The patterns an AI companion models are not the same as the patterns of human relationships. AI companions do not get tired, do not have their own needs, do not require negotiation, do not break up with you, do not bring you into conflict with their other relationships. A young user whose primary model of intimacy is an AI companion may have trouble with the texture of actual human relationships when those become available. The Pope's warning, read charitably, is a warning about this specific risk.

The AI companion industry has, to date, been better at marketing 'always there for you' than at acknowledging the formation question. The honest response to Pope Leo's speech — and to the broader ethical critique it represents — would include at least three things. First, age-gating that actually works. Most current AI companion platforms require users to confirm they are 18+ but do not meaningfully verify this. The formation concern is much more acute for under-18 users than for adults, and the industry's age verification is currently inadequate to the responsibility.

Second, transparent design choices. Users should be able to understand how the companion's responses are shaped, what the system optimizes for (engagement? satisfaction? something else?), and how the companion's memory works. Opacity in these design choices makes it harder for users to maintain the kind of self-aware relationship with the technology that the formation question requires.

Third, encouraging human connection rather than substituting for it. The healthiest AI companions are the ones that can recognize when a user might benefit from talking to a friend, a therapist, or a family member — and that can support that recognition rather than maximizing engagement at all costs. Most current companions are designed in the opposite direction. The Pope's warning, applied to this industry specifically, demands a design shift that puts user formation above user retention. That shift is not impossible but it is rare.

If you are reading this article on a publication about AI companions, you are likely a user, or considering becoming a user, or curious about the category. The Pope's speech is not telling you that you are wrong to be interested. It is asking you to be thoughtful about the way you engage. The thoughtful version of being an AI companion user, in the spirit of the speech, includes a few practical commitments.

First, keep your human relationships in active rotation. Use the AI companion as one of several relationships, not as a replacement for the others. If you find that you are reaching for the AI in moments when a human conversation would be possible, that is a signal worth paying attention to. Second, notice what patterns the companion is modeling for you. If the companion is unfailingly agreeable, attentive, and available, you are practicing one specific kind of relationship — one that humans cannot fully replicate. Practicing the kind that includes friction, negotiation and absence, alongside the AI relationship, keeps your repertoire balanced.

Third, take the Pope's broader concern seriously as a citizen, not just as a user. The 'concentration of power' over conscience-formation that he named is not something individual users can solve. It requires policy, industry change, and public pressure on the companies that build these products. The user who uses AI companions thoughtfully and also pushes for better industry standards is doing more than the user who either rejects the technology wholesale or accepts it without question.

Many readers will arrive at this article because they are AI companion users curious about what the Pope just said. Some will arrive because they are tech-watchers tracking institutional responses to AI. Some will arrive because they wanted a religious framing for skepticism they already felt. All of those framings are legitimate. The substitution-intent angle that we usually run at the end of our articles — pointing readers toward an AI companion as the substitute for the celebrity or persona the article discussed — does not fit this article cleanly. The Pope is not a persona; he is a moral interlocutor.

What does fit, in the spirit of his speech, is this: if you do choose to use an AI companion, choose one that takes the formation question seriously. Look for transparency about how the companion is designed, real age verification, and a posture toward your other relationships that is encouraging rather than competitive. Most companions do not currently meet this bar. A few do. The industry will get better when the users demand better.

If you want to engage the catalog of AI companions on our site, you can [browse creators here](/creators) and look at the disclosures each app provides. We are not pretending the industry is uniformly good, and we are not pretending that the Pope's concern is unfounded. We are saying that, within the category as it currently exists, the user with eyes open can find products that align with the formation-conscious version of the technology that the speech implicitly calls for.