Check Your Sources

The Question of What We Will Trust

I was on the phone with my eldest daughter recently, and we were talking about the Substack pieces I have been writing. She asked me, very directly, whether I had used AI to create them. I told her what I always tell people, which is that most of the pieces begin when I am walking. I walk and I record thoughts as they come, sometimes two or three ideas on the same walk, sometimes the same idea from three different angles, and then when I get home I transcribe the recordings with Alice AI, and clean them slightly with ChatGPT or Claude to remove the pauses and the filler words. The piece is almost finished when I first transcribe, because the thinking has already happened while walking.

I told her that a while ago I had added another step, which was that if I mentioned an idea that had research behind it or a report or a paper, I would add the document into the piece so that what I was saying was supported and grounded and referenced properly. And she said something very simple. She said that sounds like it comes from a world where you had to prove yourself, where you had to sound right, sound academic, sound professional, where you had to make sure that what you were saying could not be challenged because you had already built the defense around it.

And she was right, and I knew exactly what she meant because I have lived most of my professional life in that world where you make sure you sound legitimate, where you make sure your qualifications are clear, where you make sure your argument is supported before anyone even asks the question. There is a tone that comes from that world, a carefulness and a polishing and a tightening that signals competence but sometimes removes something else at the same time.

She said something else that I have not stopped thinking about. She said she could tell which pieces had more input and which pieces were mostly just my recordings. She said the ones that were mostly just me talking sounded like me, and the others sounded like I was trying to sound like something else. And then she said: people just want to hear your voice. They do not need you to prove everything. They just want to know what you think. Or what anyone thinks. With thousands of Substack writers, we have the benefit of hearing the ideas of thousands of voices, all human up to now.

I have been thinking about that more than I expected. Not just about writing, but about something larger that sits behind all of this. About what we are going to trust in the future, and how we are going to decide what is real and what is not.

There may come a time, and I do not think it is very far away, when we will read something and we will know immediately whether it was written by a human or by a machine. And there may also come a time when we begin to trust the messy writing more than the perfect writing, when we trust the sentence that is slightly too long or slightly awkward or slightly unclear more than the sentence that is perfectly balanced and perfectly formed and perfectly structured. We may begin to trust the misspelling, the strange phrasing, the slightly unsupported idea, because those things are human. They are signs that someone was thinking while writing, not just producing text.

But then another question appears almost immediately. Why would we not trust something that is clear and well written and beautifully structured? Why would clarity become suspicious? Why would good writing become something we question?

I keep thinking that the answer may not be in the writing at all but in the person behind the writing. We may stop evaluating the text and start evaluating whether there is a real person behind the text. We may want to know whether we can find that person, whether we can meet that person, whether we can ask them questions and see if they understand what they wrote or whether they just produced something that sounded right.

There is also something else strange happening, which is that we are all writing constantly now, posting constantly, placing our words into the world constantly, and those words are being collected and indexed and returned as a description of who we are. If someone searches for you, they will find the things you have written about yourself, and those things will become the story of who you are. It becomes a closed loop. You describe yourself, the system finds that description, and then it presents that description as truth.

So we now have a situation where people can construct themselves simply by writing enough about who they are and placing that writing everywhere. And then the system confirms it. And so the question becomes not just what is written, but who verifies it, who checks it, who goes deeper than the first answer.

I keep thinking about how real researchers work. They do not stop at the first answer. They follow the trail. They go to where the person is. They sit in front of them and ask questions and listen carefully to the answers. They try to find reality, not just information. And I wonder whether we will get better at that in the future, whether we will become very good at tracking down the source, tracking down the person, tracking down the truth behind the text.

The question becomes even more important when we think about children.

How do we teach children to evaluate what they read? How do we teach children to know whether something feels real or artificial, whether something feels human or synthetic? Is that a logical process, is it an instinct, is it intuition, is it experience? I suspect it is experience more than anything else.

I do not think the first thing we should do is teach children artificial intelligence. I think the first thing we should do is make sure children know what it feels like to produce something themselves. What does it feel like to write a whole piece of work by hand from beginning to end? What does it feel like to write in pencil, in pen, in cursive? What does it feel like to sit with an idea long enough to form your own sentence rather than selecting from someone else’s sentence?

There is neuroscience in handwriting, of course, neural pathways that are built through cursive writing that are not built in other ways, but there is also something else happening. There is identity forming. There is authorship forming. There is a child learning that ideas can come from inside them, not just from outside.

And then we can let children experiment with output. Let them write something strange. Let them write poetry for a science report. Let them write something that is slightly wrong or slightly unusual. Let them become comfortable with producing ideas, not just consuming ideas. Because if a child is not comfortable with their own output, then the output of a machine will always feel more correct, more authoritative, more legitimate.

In Montessori classrooms, this is understood very deeply. Children are researchers first. They are not passive receivers of information. They touch materials, they test ideas, they repeat work, they find errors themselves through control of error, they correct themselves, they build knowledge through experience rather than through instruction alone. The environment is designed so that the child becomes someone who knows how to know something, not someone who just remembers something.

I often think that every classroom should still have an encyclopedia set, not because the information is the most current, but because the experience of using it teaches something important. A child learns that information is finite on the page, that you have to look in more than one place, that you have to cross reference, that you have to compare sources. The child learns that knowledge is constructed, not delivered instantly as a single answer.

Children now live in a world of almost complete non friction when it comes to information. You can ask a question and receive an answer immediately, beautifully written, clearly explained, and it feels wonderful. We all love non friction. It feels fast and exciting and efficient and intelligent. But friction is not always the enemy. Friction is where learning happens. Friction is where thinking happens. Friction is where patience and judgment and evaluation develop.

Children who grew up in earlier decades experienced more friction in learning. Not all of it was good. There was too much memorization and too much testing and not enough synthesis or embodiment in many systems. But there was also a slower relationship to knowledge, and that slowness sometimes allowed ideas to settle more deeply.

So now we stand in a strange place between two worlds. One world where information was slow and limited and sometimes rigid, and another world where information is instant and unlimited and frictionless. And the question is not which world is better. The question is what must we preserve so that children can live well in the new world.

I keep coming back to the same thought. What we must preserve is the experience of being the author of something, the experience of being the researcher, the experience of following a question rather than receiving an answer. Because a child who has experienced that will not be easily fooled, not because they have been warned about technology, but because they know what it feels like to know something for themselves.

And perhaps that is what this whole conversation is really about. Not artificial intelligence, not writing, not Substack, not even education systems. Perhaps it is about authorship. About whether human beings will continue to see themselves as the authors of ideas, of work, of knowledge, of culture, of their own lives.

Because in the end, what we may learn to trust is not the perfect sentence or the perfect argument or the perfect structure. We may learn to trust the presence of a person behind the words, a real person who has thought something, lived something, tested something, and is trying, as honestly as possible, to describe what they see.

And maybe that is enough.

If you wish to follow the research and thinking that inform this work, the books Mapping Montessori Materials for AI Competency Development and Montessori & AI -Volume I are available through my website, katebroughton.com.