The Preschoolers Were Not Raw Material

The woman in front of me at the laundromat was folding tiny socks.

Pink ones. Blue ones. One with a cartoon shark on it. She kept checking the dryer like it had swallowed a child instead of a blanket, and every few minutes she looked down at her phone and made the face people make when the world has asked them to sign something rotten.

I know that face.

I made it at the post office when management handed us a new form with six boxes, three lies, and a sentence at the bottom saying we understood everything. I made it in hospitals. I made it in rented rooms with landlords who believed a clause was the same thing as a conscience.

The woman jabbed at the screen with her thumb and muttered, “They want to film them now.”

Nobody asked who.

Everybody knew.

There is always a they.

This time they wanted preschool teachers wearing cameras. Little classroom Panopticons strapped to adult bodies, catching the approximate first-person view of children being children. Blocks. Crayons. Runny noses. Small hands learning how to hold the world without dropping it.

The footage would be used to train AI models for assessing classroom interaction quality.

There it is. That dead language again. Classroom interaction quality. Three words wearing lab coats so nobody notices the thing underneath them breathing.

A child laughs. A child cries because the red cup is gone. A child tells a teacher his mother was yelling this morning. A child falls asleep during story time with his mouth open and his whole defenseless face turned toward the ceiling.

The machine calls this data.

The researchers called it normal interactions during regular classroom activities. That is how you make a theft sound like weather. Normal. Regular. Approximate. Processed. Cloud-based. Whenever possible.

Whenever possible is a beautiful little sewer pipe of a phrase.

Faces and names would be censored whenever possible. Data might be processed by cloud-based AI services. The documents did not seem interested in saying which services, which models, which companies, how long the footage would live, where it would sleep at night, who would be allowed to take it out for a walk.

Just sign here, or don’t sign, or fail to understand that not signing is somehow a yes.

That part got me.

The automatic yes.

Unless the parents formally opted out, the kids were in. Tiny citizens of the future, enrolled by silence. Maybe the form sat in a backpack under a banana peel. Maybe it came home in English to a house where English is the language of bills and fear, not bedtime. Maybe a mother working the late shift saw it at midnight with aching feet and thought, Christ, another school paper, and put it beside the coupons.

The machine does not care how consent got lost.

It just hears yes.

I have spent a lifetime watching powerful people build elaborate machines for turning silence into permission. Bosses do it. Governments do it. Lovers do it when they are cowards. Tech people have simply automated the old trick and added better lighting.

The parents revolted.

Good.

There are still moments when the animal in us remembers what the committee forgot. A parent sees a camera pointed at a preschooler and does not think, ah, innovation. She thinks, why is that lens near my child? She thinks, who gets the footage? She thinks, why are the words so soft if the thing is so harmless?

Soft words are where the knives hide.

The study died early. The university said it had terminated the thing after feedback from community partners, which is how institutions say the villagers showed up with torches without admitting there were torches or villagers.

I like parents when they are angry for the right reasons. Not the book-banning tantrums. Not the televised culture-war theater with hair spray and slogans. I mean the old mammal fury. The hand across the doorway. The no that comes before the spreadsheet.

Because preschool is not a mine.

That should not need saying, but here we are, saying it in the year of our Lord 2026 while men with degrees and grant money try to convince everyone that children sitting cross-legged on a rug are an untapped resource.

AI is hungry. That is the part they keep dressing up.

It needs books, paintings, voices, faces, medical records, driving patterns, homework, search histories, street photos, music, arguments, mistakes, confessions, and now maybe a four-year-old trying to pronounce chrysalis. The hunger is never introduced as hunger. It arrives as research. Safety. Assessment. Personalization. Quality.

Quality is another word that has been beaten half to death in a conference room.

I do not doubt that some teacher somewhere could use help. Teachers are tired in a way that makes ordinary tired look like a vacation. They buy supplies with their own money. They get blamed by politicians, parents, administrators, and children who have not yet learned that biting is frowned upon. If someone built a tool that gave a teacher back ten minutes and did not ask for a child’s face in exchange, I would buy that someone a drink.

But that is not the smell coming off this thing.

This smells like the new frontier, which is always somebody else’s backyard. The gold rush with smaller boots. The railroad through the nursery. Men and women who speak carefully about ethics while looking for more raw material to feed into the furnace.

And the funny thing is, children are exactly what the machine cannot understand.

It can classify their gestures. It can measure teacher response time. It can mark warmth, engagement, compliance, whatever the grant proposal promised. It can turn a classroom into numbers and the numbers into a dashboard and the dashboard into a meeting where everyone nods because nobody wants to be the barbarian who questions progress.

But it cannot know what it means when a child trusts you.

Trust is not a metric. It is not approximate. It is not cloud-based. It is built in crumbs and glances and the strange holy patience of tying the same shoe eleven times before lunch.

A camera sees the shoe.

It does not see the patience.

A model sees a pattern.

It does not know the teacher skipped breakfast. It does not know the kid with paint on his sleeve has not slept right since his father left. It does not know that the migrant mother who could not parse the consent form has already signed too many things she did not understand because America keeps handing her paper and expecting gratitude.

That was the detail that stuck under my fingernail. The mother who spoke English and still found the language vague, wondering what happened to the families who did not have that advantage.

There is your human angle, if anyone still remembers humans.

The people most likely to be harvested are always the people least equipped to refuse. Poor people. Immigrants. Children. Patients. Workers with badges that stop opening doors when they get difficult. The future is very brave when it comes to experimenting on people with no lawyers.

I can already hear the defense. Nobody meant harm. The researchers wanted to improve classrooms. The data would be handled carefully. The parents misunderstood. The backlash was unfortunate.

Maybe.

I have known plenty of harm that arrived with good posture and a clean shirt.

Intentions are cheap. Outcomes charge interest.

If you want to point cameras at children for the benefit of machines, every parent should know exactly what will happen, in language a tired person can understand while standing over a sink full of dishes. The default should be no. Not silence. Not confusion. Not a backpack lottery. No.

Make the machine beg for permission like everybody else.

The socks in the laundromat came out warm. The woman folded them slowly, pressing each pair flat like she was trying to restore order to one small corner of a busted planet. Then she put the phone face down and zipped the bag.

Outside, the evening traffic dragged itself along, red lights blinking in the wet street.

Somewhere, a classroom was still just a classroom.

For now.

Source: Parents Explode in Fury at School’s Plan to Constantly Film Their Children to Train AI