We recently had the opportunity to host Dr. Miguel Cardona, former U.S. Secretary of Education, for a GoTeacher conversation about AI, student safety, and what school leaders should be thinking about as these tools spread across classrooms.
Cardona has spent his career in public education, from the classroom to leading the U.S. Department of Education. We wanted his perspective on a question we’re hearing more and more from district leaders: How do schools embrace what AI makes possible without repeating the mistakes we made with earlier waves of technology?
One part of his argument has stuck with me.
Secretary Cardona described the last 25 years of technology in education as three chapters.
The first was the internet.
Schools embraced it because suddenly students could access information and experiences that had previously been unavailable to them. A student who might never visit the Smithsonian could explore it from a classroom. The world opened up.
But we didn’t fully understand the risks yet, and the safeguards came later.
Then came social media.
Again, there was real value. Students could communicate and build community in completely new ways. Cardona talked about watching his own children use social media and online games to stay connected with friends during the isolation of COVID.
But again, we were late to the risks.
Now we’re entering chapter three: artificial intelligence.
And his argument was simple: we have a chance not to repeat the same pattern.
AI Is Already in Your District
One thing districts need to accept is that this isn’t a future conversation.
AI is already inside schools.
Teachers are using it to save time. Students are using it for assignments. Existing EdTech products are adding AI features. New applications are appearing faster than districts can review them.
Cardona made the point very simply:
“They’re doing it with or without us.”
That changes the question.
The decision is no longer whether we introduce students to AI. For many students, that introduction already happened.
The question is whether schools are going to help shape how they use it.
The Risk Isn’t Theoretical
During the conversation, Cardona referenced something we had previously shared with him from our work with a school district.
When the district looked at the AI applications students were accessing through district-issued devices, one of the most heavily used categories included AI romance and companion tools.
Middle and high school students were having conversations and forming relationships with artificial characters.
Cardona’s reaction was immediate.
“When I heard that, it shook me because it made me feel like, damn it, are we losing an opportunity here to protect our kids?”
That example gets at why I think this conversation needs more urgency.
Most district AI conversations I’ve been part of start with cheating.
Can students use ChatGPT for an essay? How do we know whether something was AI-generated? What should the acceptable-use policy say?
Those are reasonable questions.
But AI creates a much broader student-safety problem.
Students can now interact privately and continuously with systems designed to respond like humans. They can share personal information with tools the district may not know they’re using. They can authorize applications to access district accounts and data without understanding what permissions they’re granting.
And the technology keeps changing.
A district can review an application today and discover six months later that the product has introduced entirely new AI functionality.
This is no longer just an academic-integrity issue. It is becoming part of the district’s responsibility for student safety.
Blocking Everything Isn’t the Answer Either
The obvious response would be to lock everything down.
That’s not the argument Cardona made.
Students are going to graduate into a world where AI is part of how people work, create, research, communicate, and solve problems.
Avoiding it doesn’t prepare them for that world.
Cardona talked about his son, who is entering music production and already thinking about how AI is changing that profession. The same thing is happening across almost every field our students will eventually enter.
If schools simply wait until AI is “figured out,” our students will be waiting while the rest of the world keeps moving.
But embracing AI without understanding what’s happening isn’t leadership either.
That’s the balance districts have to find.
As Cardona put it, we need “a healthy dose of optimism of what’s possible with a healthy dose of skepticism.”
There Is No AI Playbook
When the conversation turned to what school leaders should actually do, Cardona’s answer wasn’t a checklist.
He compared this moment to leading schools during COVID.
Six years ago, superintendents weren’t infectious-disease experts. They suddenly had to understand ventilation, transmission, testing, public health guidance, and dozens of things that had never been part of running a school district.
They learned. They found experts. They worked together. They changed their systems as new information emerged.
AI leadership is going to require something similar.
“You’re not expected to be an expert in AI to lead it.”
I think IT directors and superintendents need to hear that.
There is a temptation right now to wait for the perfect AI policy, the perfect state guidance, or the perfect set of approved tools.
It’s not coming.
Whatever policy you write today will probably need to change. Whatever tools you approve will evolve. New risks will appear. New opportunities will appear.
The goal isn’t to finish your AI strategy. The goal is to build a district capable of continuously adapting it.
Start by Knowing What Is Actually Happening
There is one practical place I think every district can start: get visibility.
Before debating which AI tools should be approved next year, understand what’s happening this year.
What are students already accessing? What are teachers already using? Which applications have access to student information? Which have appropriate privacy agreements? Which applications have introduced AI features since they were originally approved?
And while you’re doing that, look at what you’re already paying for.
Cardona made a point that will be familiar to almost every district leader:
“I would argue that every district on this call right now has programs that are being paid for that are not being used.”
AI shouldn’t simply become another layer of software spending.
Districts have an opportunity to use this moment to look critically at the technology they already have, understand what’s actually being used, remove what no longer provides value, and make room for what comes next.
Trust Requires Admitting We Don’t Have Every Answer
One of my favorite moments came when Cardona was asked how districts can build trust with families around AI.
His answer was:
“Don’t act like you know it all.”
That’s probably some of the best AI leadership advice I’ve heard.
Parents don’t need their superintendent to pretend to know exactly what AI will look like five years from now. Nobody does.
They need to know that the district is paying attention.
Tell families which tools you’re exploring. Explain why you think they could improve learning. Be honest about the risks. Show them the safeguards you’re putting in place. Ask for input.
Most importantly, make it clear that protecting students and preparing them for the future are not competing goals.
We have to do both.
We Don’t Have to Be Late Again
The internet gave students access to information at a scale we had never experienced.
Social media gave them access to people at a scale we had never experienced.
AI is giving them access to something different: systems that can communicate, create, advise, persuade, and increasingly behave like another person.
We shouldn’t panic about that, but we shouldn’t be casual about it either.
We’ve already seen what happens when technology moves into children’s lives faster than the adults responsible for them understand it. This time, we have the benefit of experience.
We know adoption will move faster than policy. We know students will find tools adults don’t know about. We know blocking everything won’t work. And we know waiting until something goes wrong isn’t a strategy.
Cardona said something near the end of our webinar that captures the challenge well:
“I don’t want big tech telling me what schools should look like. I want educators who understand children, who understand families, who understand leadership and how systems work, to be driving that.”
That, to me, is the bigger lesson from having Secretary Cardona join us.
AI is going to influence what education becomes. Schools shouldn’t wait for technology companies to decide what that looks like for them.
Educators need to lead it.
We were late with the guardrails before. We don’t have to be late again.
