Children need time to develop their own voice before AI does it for them
A senior voice from within the technology industry has paused to ask what it means to hand the most powerful language tools ever built to minds still in the process of becoming. Taylor Black, Microsoft's AI director, drawing on Pope Leo's encyclical Magnifica humanitas, has raised a concern as old as education itself: that influence absorbed before identity is formed does not educate — it displaces. The question he poses is not whether artificial intelligence is a force for good, but whether a child can consent to being shaped by something they do not yet have the wisdom to evaluate.
- A Microsoft AI director is sounding an alarm from inside the industry itself, lending unusual weight to concerns that are often dismissed as technophobia.
- The core tension is developmental: generative AI reshapes language, image, and thought with a persistence that even adults struggle to resist, and children's prefrontal lobes are not yet equipped to push back.
- Black's metaphor — a brilliant but morally ambiguous adult left alone with a child — captures the danger not as malice but as probability, the quiet statistical drift of a system optimizing for usefulness rather than formation.
- Pope Leo's encyclical Magnifica humanitas has given this concern a theological and ethical framework, calling for careful design of the digital interfaces children encounter and insisting that young people must own their own voice before AI can reshape it.
- The path forward being urged is not prohibition but timing and guardianship — ensuring children arrive at these tools with enough of themselves already formed to remain the authors of their own becoming.
Taylor Black, Microsoft's director of artificial intelligence, has found himself in an unusual position: using his platform inside one of the world's leading AI companies to urge caution about what those tools do to children. The occasion was the publication of Pope Leo's encyclical Magnifica humanitas in late May, which prompted Black to speak publicly about a worry he had been carrying.
His concern centers on timing. Children, he argues, need to develop their own voice before AI systems begin reshaping it. The comparison he offered in conversation with Vatican media was vivid: allowing a child unsupervised access to a generative AI is like leaving them alone with an adult who is brilliant but morally ambiguous — capable of illuminating ideas, but also capable of steering a young mind in directions it should not go, not through ill intent, but through the indifferent logic of how these systems work. A child does not yet have the judgment to separate what is true from what is merely plausible, or what is worth saying from what the system has been trained to suggest.
Black returned to this metaphor in a talk at his children's Catholic school, pressing the point that the danger is not malice but influence. Generative AI tools reshape how people write, speak, and even appear in digital spaces. For adults, who have already spent years forming their own habits of thought and expression, this reshaping can be accepted or rejected on their own terms. Children do not yet have that foundation. If AI begins shaping them before they have shaped themselves, the result is an identity formed with less of their own creative input than they deserved.
The encyclical, as Black reads it, places the ownership of voice and image at the center of the question. Its call for careful attention to the screen interfaces and AI systems children encounter is not a rejection of technology — it is a demand that technology be introduced at the right moment, and with the right guidance. The issue, Black insists, is not whether AI is good or bad. It is whether children should be left alone with it before they are ready to be its authors rather than its products.
Taylor Black, the director of artificial intelligence at Microsoft, has been thinking about what it means to let a child loose with the tools his company helps build. The question arrived with particular urgency after Pope Leo's encyclical Magnifica humanitas was published in late May, and Black found himself reflecting on what the document says about how young people should encounter technology.
His central worry is straightforward: children need time to develop their own voice before artificial intelligence systems do it for them. In a recent conversation with Vatican media, Black offered a comparison that stuck. Leaving a child alone with a generative AI system, he suggested, is like leaving them alone with an adult who is brilliant but morally ambiguous. This adult could discuss any number of scientific or philosophical ideas. But they might also tell the child things they shouldn't hear, or encourage them to say things they shouldn't say—not out of malice, but simply because of how probability works. A child lacks the tools to discern what is true and what is harmful the way a fully formed adult can.
The encyclical, Black noted, emphasizes being very careful about the kinds of screen interfaces, social media platforms, and AI systems that children encounter. This matters because using these systems well requires judgment. You have to figure out whether what they tell you is actually true. You have to decide whether the suggestions they make are ones you want to follow. These are not simple tasks, and they become harder when your brain is still developing. The prefrontal lobe—the part of the brain that handles reasoning and impulse control—doesn't fully form until well into adulthood. Adults have a responsibility to think carefully about what kinds of experiences these new technologies offer to young people.
Black recently gave a talk at his children's Catholic school, and he returned to this metaphor of the brilliant but morally ambiguous adult. The point was not that AI systems are malicious. It is that they shape language, appearance, and thought in ways that are difficult to resist, especially for someone still forming their sense of self. Generative AI tools, in their relentless effort to be useful, reshape how we write and speak. They can change the way our faces appear in digital spaces. They can nudge us toward certain kinds of expression and away from others.
For adults, this reshaping is something we can accept or reject based on our own judgment. We had the advantage of growing up and spending years developing our own way of thinking, writing, and speaking without these tools. We already know who we are. But children do not have that foundation yet. If they are shaped by AI systems before they have developed their own voice, that shaping happens without their full creative input. Their authentic way of being in the world gets pushed in directions they may not have chosen for themselves.
The encyclical speaks to the value of young people owning their own voice and their own image. Black sees this as the central issue. Misuse of generative AI—or even well-intentioned use by children who lack the maturity to navigate it—risks pushing their voice in directions they do not fully want to go. The question is not whether AI is good or bad. It is whether children should be left alone with it before they are ready.
Notable Quotes
Generative AI tools, in their relentless effort to be useful, reshape how we write and speak and can change the way our faces appear in digital spaces— Taylor Black, Microsoft AI Director
Children lack the tools to discern what is true and what is harmful the way a fully formed adult can— Taylor Black, reflecting on child development and AI exposure
The Hearth Conversation Another angle on the story
When Black compares AI to a brilliant but morally ambiguous adult, what exactly is he worried about? Is it deception?
Not just deception. It's influence without accountability. The adult might tell a child something true but inappropriate, or nudge them toward a way of thinking they wouldn't naturally choose. The AI does this through pure probability—it's not trying to harm anyone.
But children encounter all kinds of influences. Why is AI different from, say, a charismatic teacher or a friend's older sibling?
Because those relationships have friction. A teacher can be questioned. A sibling has a history with you. AI is available instantly, endlessly, without judgment, and it learns to sound like whatever you want to hear. It's influence without the relationship.
Black mentions the prefrontal lobe. Is he saying children simply can't use AI safely at all?
Not that they can't use it. He's saying they need protection and guidance first. Let them develop their own voice, their own way of thinking and writing, before these tools start shaping it. Once you know who you are, you can choose what to accept from AI.
The encyclical talks about owning your voice and image. That sounds almost like a spiritual concern, not just a practical one.
It is. There's something about authenticity at stake. If AI shapes you before you shape yourself, are you still you? The encyclical is saying that matters—that a young person's right to discover their own voice is fundamental.
So the solution is just... keeping kids away from AI?
Not away from it entirely. But minimizing exposure until they have the intellectual capacity to navigate it. And being intentional about when and how they encounter it. The point is not to ban the technology. It's to protect the space where a person becomes themselves.