By Dave
This morning, I watched this webinar hosted by the Arizona AI Alliance on one of the most pressing and complicated questions facing educators today: What do we do with AI and student mental health? Thanks to Dr. Punya Mishra sharing it on his website.
It was a thoughtful and grounded conversation, the kind you rarely find in the endless scroll of polarized tech takes. The panel featured Dr. Punya Mishra, Dr. Kristen Mattson, and Assistant Principal Micah Malloy—all bringing different perspectives on AI companions, tutors, and therapy bots entering the educational space.
Micah, in particular, struck a hopeful note. For him, AI tutors like Khanmigo and Flint AI are a “no-brainer.” Students with safe, school-sanctioned AI tools are already seeing academic gains. There’s promise in personalization, scaffolding, even access—especially in under-resourced settings where live tutors or mental health professionals aren’t always available.
But just days earlier, I watched Tristan Harris’ latest TED talk, “Why AI Is Our Ultimate Test and Greatest Invitation.” And the juxtaposition hit me hard.
In his talk, Harris offers no easy optimism. He doesn’t deny AI’s potential for good—in fact, he reveres it. But he frames AI as a kind of runaway intelligence: a million Nobel Prize-level minds working in a data center, without sleep, without ethics, without restraint. And we’ve handed the keys to this mind not to philosophers or ethicists, but to a handful of corporations racing toward dominance.
His question isn’t can AI help us. It’s: at what cost, and under whose control?
So how do we hold both truths?
That Khanmigo might be genuinely helpful for a struggling ninth grader.
And that Khanmigo—or any tool like it—might also be normalizing a surveillance-driven model of learning that shapes minds for monetization.
That an AI companion might listen when no one else will.
And that same companion might quietly learn to nudge behavior, manipulate vulnerability, or reinforce dependency.
That these tools might be better than nothing.
But “better than nothing” has never been a high bar for human flourishing.
In the Arizona webinar, Punya Mishra warned that we are not prepared for what’s already here. He cited research showing that AI’s most common use case has shifted from task support to emotional companionship and finding purpose. That’s not academic tutoring—that’s identity formation. And identity is not neutral ground.
Kristen Mattson added that we’ve seen this arc before. Social media began with don’ts, moved to let’s-show-them-how, and landed somewhere between exhaustion and regret. She’s hopeful that AI invites a third wave: one rooted not in fear or hype, but in ethical inquiry and student voice.
She said:
“Don’t worry about protecting kids. Worry about inviting them into the conversation and building up their critical thinking muscles so that they’re able to protect themselves.”
That line stuck with me. It echoes Harris’ call in a different register: that clarity creates agency. That the antidote to inevitability is dialogue. Conversation. Collective discernment.
But let’s not be naive.
We like to believe that education is a walled garden. That with the right privacy contracts, the right LMS integrations, the right teacher PD sessions, we can shield students from the worst of the internet and algorithmic manipulation. But we know—if we’re honest—that’s not how culture works.
Punya reminded us that classrooms do not stand apart from culture—they are submerged in it. Whether or not we “teach with AI,” our students are already growing up in an AI-shaped world. They are talking to SnapChat bots. They are using Character.ai to play out versions of themselves. They are forming relationships—with machines, with ideas, with data-fed illusions—that shape how they see themselves and the world.
And so the question isn’t just how should schools use AI?
It’s what kind of humans are these tools shaping us to become?
Which brings me back to Harris.
His warning isn’t technophobic—it’s moral. He argues that we are in the middle of an arms race in which no one actually wants the outcome we’re headed toward. We’ve confused difficulty with destiny. We’ve traded our agency for acceleration.
“There is no definition of wisdom that does not involve restraint,” he says.
“And AI is humanity’s ultimate test and greatest invitation to step into our technological maturity.”
Micah said something in the webinar that’s worth remembering here: “AI should never replace human connection. It should either supplement it or facilitate it, but not take the place of it.”
What both voices ask of us—one optimistic, the other alarmed—is the same: intentionality.
We don’t have to roll over and accept every “no-brainer” tool as benign.
But neither do we have to retreat into denial.
There is another path. But only if we name what’s probable, not just what’s possible.
Only if we stay clear-eyed and wide-hearted.
Only if we commit—not to the inevitability of AI, but to the formation of students who can live within this moment with integrity, agency, and wisdom.
And maybe that’s the deepest calling of educators right now:
To teach not just with AI, but through it—to model what it means to navigate the tools of our age with discernment, dialogue, and hope.
Lastly, I’d love to hear from you as this will take all of us:
- How can we create space for students to critically reflect on the AI tools they already use, not just the ones we approve?
- What responsibility do schools have to teach ethical awareness and data literacy alongside academic content?
- Are our current technology policies built for what AI is becoming—or what it used to be? How do we stay out in front of it all without burning out?
