The Chair symbolThe Missing Seat

TMS EDITORIAL 045

Who Is Programming Our Children’s Future?

Artificial intelligence is entering schools before children have a reliable right to know, challenge or escape what the machine decides about them.

Seat Affected: A child’s right to an open future

Seat Status: At risk of automated judgment

Location: United States, with global implications

Theme: Science & Innovation • Children & Families • Education

A child’s school record should describe part of a childhood.

It should not become a permanent prediction of who that child is allowed to become.

Artificial intelligence is entering education through tutoring, writing assistance, plagiarism detection, behavioral monitoring, threat assessment, attendance intervention, special-education support and systems designed to predict which students may struggle.

Some tools may help teachers notice needs earlier and give students individualized support. Refusing every new tool would not protect children from underfunded schools, overworked teachers or human bias.

But adoption is moving faster than accountability.

Who trained the system? Which children were represented in the data? What outcome was it optimized to produce? Can a parent see the record? Can a child challenge a label? How long is the inference stored? Is it sold, shared or used to improve a commercial product? Does the judgment follow the child to another school?

The danger affects all children. Black children reveal why the danger is not theoretical.

The Government Accountability Office found that Black girls received more frequent and more severe school discipline than other girls. In 2017–18, Black girls were 15 percent of girls in public schools but received almost half of suspensions and expulsions. GAO also found harsher punishment than white girls for similar infractions.

Now imagine training a risk model on those records.

The machine does not need a field labeled “race” to learn America’s racial patterns. Neighborhood, attendance, discipline history, language, school resources, family income and prior referrals can act as proxies. Historical judgment enters the system as data. The output returns as a score that looks objective because a computer produced it.

That is automation’s most convincing disguise: prejudice can lose its human voice without losing its effect.

UNICEF warns that generative AI inherits known problems of bias and opacity and may add unpredictable outputs. Its child-data work asks a deeper question: who controls children’s information, who benefits from it and whether children have a voice in systems shaping their lives.

Schools should not wait for a scandal. Before an AI system affects opportunity, discipline, safety assessment or access to services, children need enforceable protections:

  1. Notice that AI is being used.
  2. A plain-language explanation of what it does.
  3. Access to the data and significant inferences.
  4. A human decision-maker with authority to disagree.
  5. A meaningful appeal.
  6. Testing across race, sex, disability, language and income.
  7. Strict retention and deletion rules.
  8. A prohibition on selling or repurposing student data without genuine consent and legal authority.
  9. Public reporting of errors and disparate outcomes.
  10. A rule that no high-stakes decision rests solely on an automated score.

Children are not defective adults. They change rapidly. They make mistakes. They recover. They surprise us. A prediction can become a cage when teachers, counselors and institutions begin treating it as destiny.

The purpose of education is to expand possibility, not automate yesterday’s expectations.

Return to the Seat

TMS will ask school districts which AI systems they use, what data those systems collect, whether civil-rights testing occurred, and how families can appeal.

If human institutions already perceive and punish Black children differently, what happens when those judgments become training data, risk scores and automated recommendations?

Who is missing? Ally = Action. Take Your Seat.

Sources and Receipts

Verification Note

The editorial describes documented risks and asks prospective accountability questions. It does not claim that every school AI system discriminates or that every listed use is deployed in every district.

← Return to all editorials