The Cosmic Vision & Human Evolution
When we present the world to the elementary child, we start with everything, the whole universe. We start with a story of origins, beginning with the Big Bang. Through the story, we explain how a number of laws regulated all particles in the universe: laws of physics, chemistry, and so on. We show how agents, living or non-living, played a role in making the emergence of our solar system possible, our planet, and all life on it. These great stories of the universe and the coming of life help us realize how lucky we are to exist today, and how lucky we are to have all that this world has to offer. All of this is possible because the world was given laws and opportunities to realize beauty and complexity, through a long and fascinating spontaneous work of adaptation.
We carry the same perspective when we think about human civilization. Again, we have origin stories, including the origin of humanity. We ask: where did we come from? What was special about us when our species evolved? We speak of the special gifts of human beings, the mind to think and imagine, the hands to work and create, and the heart to love. We look at the evolution of human civilization as another process of adaptation, but a creative one. Our special gifts allow us to modify the planet and our environment.
This creative capacity is not only physical but also intellectual. A child, in their first years, constructs their intelligence through interaction with the environment and through the inheritance of human culture. Thanks to this work of the child, generation after generation, our collective intelligence grows. We inherit what has been created before us. Some inventions, such as language, oral first, then written, have empowered this expansion of knowledge. Written language ensured that knowledge could be passed on not just to the next generation, but far beyond. Writing allowed us to formalize, deepen, and refine our understanding.
So, over time, humanity has created tools to expand our intelligence. The first humans were limited to what they could experience and what they could communicate orally. Today, our intellectual capacity builds upon thousands of years of evolution. Technologies like the internet have further accelerated this expansion. They multiplied access to collective knowledge, removing the limits of direct physical interaction. Yet, frontiers still remain, language barriers, cultural silos, inaccessible knowledge. Even so, we continue inventing technologies to bridge those divides.
AI as Extension of the Supramind
And now, we have artificial intelligence. A technology that can do things that were science fiction not so long ago. I can talk to a computer, and it will interpret my words and respond. If it understands me, statistically or mathematically, and has access to vast knowledge, and even some basic reasoning, it’s almost as if I’m speaking to someone with access to everything on the internet. Like writing or the internet, AI has the potential to expand our access to collective intelligence. Montessori spoke of this as the “supramind”, the collective knowledge created through the years, which forms part of “supranature”, the body of human-made contributions to our world through creative adaptation.
Human beings are not limited to a single way of living. Our minds are shaped in early childhood along certain lines, through sensitive periods and other developmental phenomena. Thanks to this, we’ve transformed the world. Supranature is the outcome of this transformation. The supramind, as an extension of cognition and memory, includes all the books in our libraries, tools that expand our intelligence. In the same way humans once used tools to extend their physical power, AI extends our mental processing.
Montessori compared the evolution of civilization to the development of the fetus. Organs develop independently until the circulatory and nervous systems unify them into an organism. Similarly, civilizations developed in cultural “bubbles”, shaped by their environments. But now, with communication and transportation technologies, we are forming a global organism. Montessori called this idea “Nazione Unica”, the idea that we already are one great nation. The problem is not that we’re too separate. The problem is that we don’t realize we’re already united. We learn about differences, but we don’t learn to see our connections.
AI could support this unity. Large language models, like ChatGPT, can understand and respond in many major languages. I can now ask my students in Tanzania to write in Swahili, and AI can translate their reflections for me. That’s already one way it bridges cultures. But beyond translation, what if AI could help us understand how different cultures think and express meaning? What if it could one day recognize the ways in which languages shape intelligence itself? That potential gives me chills. Imagine if AI could grasp not just communication, but the diversity of cognitive structures rooted in culture, language, and experience. AI could support cooperation and understanding in ways we’ve never known, respecting differences instead of assimilating them.
Comparing AI Training to Human Development
Now, here’s a little intellectual gymnastics. I’d like to explore the similarities and differences between AI training and the self-construction of human intelligence, what we know, as Montessorians, from the four planes of development. First, how is AI trained? There is the initial data collection. AI is “fed” vast amounts of data, this is its first stage, absorbing and internalizing what it’s given.
Then comes pattern recognition, classification, and prediction. That’s model training and optimization. The AI tries to make sense of the data and organize it into something resembling meaning. After that comes refinement, using feedback. The AI tests predictions and adjusts based on whether responses are accepted or rejected. This is model validation. On top of this is an ethical framework, rules set by its creators to help mitigate risks and define limits. Finally, the AI is deployed and continues to refine itself with further feedback, just like ChatGPT asks which of two responses you prefer.
So yes, there are parallels. Human beings also absorb information, process it, and refine understanding through feedback. But there are big differences. AI doesn’t have curiosity. It doesn’t feel or choose. It follows objectives, not sensitive periods. It optimizes statistically within a defined framework, not guided by inner tendencies. Human beings are born with potentialities, orientation, imagination, morality, social life. These give direction to our development. For AI, those guiding forces are created externally: model architecture, training objectives, ethical guidelines.
In humans, nature gives us the DNA. For AI, we define the DNA. We give it the data and the aim. And if we’re not careful with this, we risk deviation. Just like a child absorbs the world around them, AI absorbs its data. If that data is biased, AI becomes biased. Early versions of ChatGPT couldn’t be released because they showed racism and gender bias. Montessorians know what happens when children are raised in environments filled with unexamined prejudice. They absorb it. That’s why we’re so careful in preparing the environment.
The Prepared Environment for AI
So what is the prepared environment of AI? Primarily, the data. In the case of large language models, that’s text, nearly the whole internet. But also, how that data is selected, filtered, cleaned. And then, the architecture, the rules coded into the AI. What are its goals? For example, a self-driving car’s AI must detect danger and decide how to avoid it. A language model aims to generate human-like responses. These goals, rules, and feedback loops form the AI’s potentialities. So again, we must ask: Who is deciding this? Who is preparing AI’s environment? What values are being embedded? Is it only commercial interests? Or are there people with a cosmic vision involved?
AI, Children, and Our Montessori Responsibility
AI is becoming an agent in our civilization. And like every new agent, we must ask what role it will play in nature and supranature. Will it support interdependence? Will it contribute to a harmonious whole? That’s the great test AI presents to humanity. Are we aware enough, ethically, morally, spiritually, to create a new kind of agent? If AI one day parallels human intelligence, will we have given it a direction aligned with the cosmic plan?
A more practical question often asked is: should we use AI with children? My response: I don’t think we can stop AI from being in children’s environments. So the question becomes: how do we prepare AI to interact with children? Today, no AI is suitable for children under 12. Why? Because they still show deviations, biases, and hallucinations. They sound confident, even when wrong. They trick adults. They’re certainly not ready for children.
But AI will soon be everywhere. Home devices like Google and Alexa are integrating AI. Robotics are becoming cheaper and more available. Children will interact with robots. So we must ask: will those robots be ready to interact in a way that respects development? We need AI that recognizes emotion. A concept called “Theory of Mind” comes from psychology, the capacity to understand that others think differently than I do. Some AI research is exploring this. Beyond that is empathy. Not just recognizing emotion, but responding to it. That requires understanding context, culture, and the lived experience behind words.
If AI is to be near children, it must be able to interact with their reality, with the uniqueness of each child.
That’s what we do as educators. We read each child’s state of mind. We don’t follow recipes. We respond to where they are. If we couldn’t do that, we’d be robots, applying generic rules.
So again, the question isn’t “Should we bring ChatGPT into the classroom?” The question is: What kind of AI is appropriate to support children in their development?
As Montessorians, we hold precious knowledge about human development. What is our responsibility here? Who should we talk to? What role can we play? Let’s not be driven by fear. Let’s ask what we can do. Progress is happening. We can’t stop it. But we can help it align with the cosmic evolution of things.
AI is perhaps our new cosmic responsibility. And if we can develop it ethically, it might become one of our greatest contributions to the universe. A bridge back to the whole. But like any powerful tool, it could divide or unite. And it is our collective responsibility to ensure it serves the cosmic plan.