Schools Are Rethinking Student AI: What Parents and Educators Should Know

As both a parent and a former teacher, I've been watching the conversation around AI in schools evolve remarkably quickly.

A few years ago, much of the discussion centered on ChatGPT and cheating.

Then the conversation shifted toward how teachers could use AI to save time, create lessons, provide feedback, or personalize instruction.

Now we're entering another phase.

Schools are beginning to ask harder questions about what happens when AI is placed directly in students' hands.

And some very large school systems are deciding they need more time, and much stronger guardrails, before moving forward.

New York City Public Schools has instituted a one-year moratorium on student-facing generative AI for students through eighth grade for the 2026–27 school year. High school students have more limited, supervised access to vetted AI programs, along with required AI-literacy instruction. Companion chatbots are prohibited across all grades.

Los Angeles Unified has gone even further for the current school year, restricting student access to generative AI platforms across all grades while the district reviews instructional uses and safeguards.

These aren't small districts experimenting on the margins.

They are the two largest public school systems in the United States.

That doesn't mean every district will follow them.

But it should get our attention.

I don't think there is one simple answer.

The more schools use AI, the more complicated the questions become.

That last question is becoming increasingly important.

New federal guidance on classroom technology encourages schools to judge technology by its instructional value, evidence, transparency, implementation, and effect on student outcomes, not simply by whether the technology is innovative.

New research evaluating AI-powered learning tools is also reinforcing the need to look at individual products rather than treating AI as a single category.

AI isn't one product.

“AI is good for students” and “AI is bad for students” aren't particularly useful conclusions.

The real question is what a particular tool does for a particular child in a particular learning environment.

For several years, schools have been under enormous pressure to figure out how to “adopt AI.”

I think we need to change that language.

Schools don't really need an AI adoption strategy. They need an AI decision-making strategy.

There may be AI tools that make a teacher's administrative work significantly easier without ever interacting with a student.

There may be tutoring tools that help an older student understand a difficult concept.

There may be accessibility tools that provide meaningful support for students with disabilities.

There may be AI systems that are appropriate for a high school classroom but not an elementary school.

And there may be products that simply don't provide enough educational benefit to justify the privacy, safety, developmental, or academic risks.

Those decisions shouldn't all be the same.

This is why the next stage of AI in education is likely to be much less about saying yes to AI or no to AI.

It's going to be about evaluating individual tools.

Another major development happened this month.

The American Federation of Teachers, United Federation of Teachers, and Microsoft announced a new National AI Safety & Privacy Standard for schools.

Among its protections, the agreement says student and educator data covered by the standard cannot be used to train AI models, schools retain control over data use and deletion, AI systems require human oversight, and families and educators should receive understandable information about how AI tools work and what protections are in place.

That matters beyond Microsoft.

It raises the bar for the questions schools should be asking every AI vendor.

Parents should be able to ask: What information does this AI tool collect about my child?

Schools should be able to answer.

Teachers should be able to ask: Is student information being used to train the AI?

Schools should know.

Administrators should be able to ask: What happens to our students' information if we stop using this product?

The vendor should have a clear answer.

Privacy policies shouldn't require parents or principals to become technology lawyers just to understand what happens to children's information.

One thing concerns me as a parent.

Many families still don't know which AI tools their children are using at school.

That's becoming a problem because AI is increasingly built into tools students already use.

A child doesn't necessarily open something labeled “AI.”

AI may appear inside search, writing software, learning platforms, productivity tools, tutoring programs, or other applications already familiar to students.

So I don't think parents need a giant list of every piece of technology their child's school uses.

But I do think we deserve clarity about the important ones.

1. Which AI tools are students allowed or encouraged to use?

Not simply, “Does the school use AI?” Ask for the actual tools.

2. How are those tools evaluated before students use them?

Who checks privacy? Who looks at age appropriateness? Who evaluates content safeguards? Who determines whether the tool actually supports learning?

3. When should students use AI, and when should they be expected to think, write, research, calculate, create, or problem-solve without it?

That may be the most important question of all.

Because the purpose of school isn't to teach children how to get answers as quickly as possible. It's to help them become people who know how to think.

I remember what it was like as a teacher when a new program, requirement, or technology entered the classroom.

Teachers are usually expected to figure it out while doing everything else.

AI makes that particularly difficult because the technology changes so quickly.

And many educators still haven't received formal guidance.

Recent national reporting has found that a large majority of teachers have not received formal guidance about how they should use AI tools in their work.

That's not a reasonable long-term strategy.

A classroom teacher shouldn't have to personally investigate a vendor's privacy policy, age requirements, model-training practices, safety systems, educational evidence, and product updates before deciding whether a student should use it.

Those decisions need institutional support.

Schools need a repeatable process.

If I were sitting with a school leadership team today, I wouldn't begin by asking: “What is our AI policy?”

I'd start one step earlier.

How do we decide whether an AI tool belongs in our school?

That decision should consider at least:

And then comes the part that is often forgotten:

Reassess it.

AI products change. Models change. Features change. Privacy practices change. Safety protections change.

A tool that was appropriate when a school evaluated it six months ago may not be exactly the same product today.

AI governance can't end when someone checks Approved on a spreadsheet.

I want to be especially clear about this.

I don't believe schools need to choose between embracing every new AI tool and banning AI entirely.

Those are easy positions.

The harder, and more responsible, work is evaluating what actually helps children learn.

There are reasons to be excited about AI.

A carefully designed tool may help explain a concept differently when a student is stuck. It may increase accessibility. It may help a teacher identify misconceptions or reduce administrative work that takes time away from students.

There are also legitimate reasons for caution.

Children need opportunities to struggle with difficult ideas.

They need to write. They need to read deeply. They need to solve problems. They need to talk to teachers and classmates.

They need to learn how to be bored, frustrated, confused, and eventually figure something out.

Those aren't inefficiencies we should automatically engineer away.

Sometimes the struggle is the learning.

I don't think the current pullback from student AI means schools have decided AI has no place in education.

I think something more important is happening.

The first wave was experimentation. The next wave is accountability.

Schools are beginning to ask AI companies for evidence.

Educators are asking for guidance.

Parents are asking for transparency.

Policymakers are asking about privacy and safeguards.

And large districts are demonstrating that simply because a technology exists doesn't mean schools have to put it in front of children immediately.

That's healthy.

But it also means families and educators need reliable ways to make sense of a technology landscape that is changing extremely quickly.

That is precisely why I created SafeGrade AI.

Parents shouldn't have to read a 30-page privacy policy to figure out whether an AI tool is appropriate for their child.

Teachers shouldn't have to become AI safety researchers.

And school administrators shouldn't have to start from zero every time another AI product appears.

SafeGrade AI independently evaluates AI tools through the questions that matter for children and schools: age appropriateness, privacy, safeguards, educational purpose, responsible use, transparency, and ongoing governance.

We aren't here to tell every school to embrace AI.

And we aren't here to tell every family to fear it.

We're here to help adults make better-informed decisions about the AI children actually use.

Because that is where I believe the conversation is headed.

And it's a conversation schools and families need to be having now.

For Parents

Ask your child's school: “Which AI tools are students currently allowed or encouraged to use, and how were those tools evaluated for age appropriateness, privacy, safety, and educational value?”

You don't need to argue for or against AI. Start by understanding what's actually happening.

For Educators

Choose one AI tool currently being used with students and ask: “What learning problem does this solve, and what evidence do we have that it solves it?”

For School Leaders

Create an inventory of student-facing AI tools currently in use.

For each one, identify: Who approved it? Which grades use it? For what purpose? What safeguards are required? When will it be reassessed?

If those answers aren't readily available, that's your starting point.