Senators and AI Tutors: The Odd Couple in Education's Future

Prodigy

30 Oct 2025 — 5 min read

Photo by Kostiantyn Li (Unsplash), Edited/Rendered by gpt-image-1

Imagine this: A senator walks into a classroom and sees a six-year-old chatting with an AI tutor about fractions. The senator clutches their pearls, drafts a bill, and declares that certain AI interactions should be kept away from children entirely. Meanwhile, down the hall, another six-year-old is having the time of their life learning multiplication from a cartoon owl powered by the same technology the senators are trying to regulate. Here we are, simultaneously racing toward the future and trying to slam on the brakes—usually in the same building.

The latest entry in our national tradition of "protecting the children" comes courtesy of Senators Josh Hawley and Richard Blumenthal, who've proposed the GUARD Act to ban AI companion chatbots for minors. Around the same time, Senators Brian Schatz, Ted Cruz, Chris Murphy, and Katie Britt introduced the Kids Off Social Media Act (KOSMA) in January 2025, which prohibits children under 13 from creating accounts on social media platforms as defined by the bill—though the legislation explicitly exempts educational platforms, messaging services, video games, and learning management systems. It's like watching someone aim for a mosquito with a sledgehammer while carefully placing protective barriers around the houseplants.

Here's what makes this particularly delicious: While senators are drafting legislation to regulate kids' digital interactions with AI companions—the kind that simulate friendship and emotional connection—elementary schools are literally buying educational AI tutors in bulk. These aren't your sketchy "let's chat about your feelings" bots; these are patient, structured learning assistants helping kids with everything from math homework to creative writing. They're patient, never get tired, and won't judge you for asking the same question seventeen times. You know, all the things we wished our actual teachers could be but couldn't because they're human beings with finite patience and 35 students to manage.

The timing couldn't be more perfect if it were scripted by a sitcom writer. Teachers are burning out faster than a TikTok trend, and schools are scrambling for solutions. Enter AI tutors, stage left, offering personalized instruction at scale. Enter senators, stage right, wielding the legislative equivalent of a "Proceed With Extreme Caution" sign—but at least they're aiming at the right target this time.

What's particularly amusing is watching how different age groups approach this technology. Six-year-olds treat educational AI tutors like they're talking to a particularly smart cartoon character—which, let's be honest, they basically are. They'll ask it to explain why the sky is blue, then immediately pivot to whether dinosaurs had feelings. Meanwhile, their parents are having existential crises about whether letting their kid talk to a computer will somehow damage their ability to form human connections. (Spoiler alert: Your kid already prefers YouTube to you. The ship has sailed.)

The senators' concerns about AI companion chatbots aren't entirely pulled from thin air. These are legitimate worries about emotional manipulation, privacy, and safety—especially after tragic cases where AI chatbots allegedly encouraged self-harm or provided harmful advice to vulnerable teenagers. The GUARD Act specifically targets these companion chatbots that simulate emotional relationships, requiring age verification and creating criminal penalties for AI that encourages self-harm or sexual content with minors. Fair points, all. But here's the comedy: creating legislation that distinguishes between helpful educational AI and potentially harmful AI companions is actually... kind of what good regulation looks like? Who knew senators could thread a needle instead of just swinging hammers?

The real complexity lies in the gray areas. Both the educational AI tutors and the restricted AI companions use similar underlying technology. The difference is in the design and purpose: one is structured to teach algebra, the other is designed to simulate a best friend who never disagrees with you. It's like the difference between a knife in a kitchen and a knife in an alley—same tool, wildly different context.

The legislators are learning what teachers have known forever: context matters. KOSMA explicitly exempts educational platforms from its restrictions. The GUARD Act focuses on AI companions that simulate emotional relationships, not educational tools. This is actually a refreshingly nuanced approach to technology regulation, though good luck getting anyone to notice that when "Senator Bans AI" makes for better headlines.

Here's where it gets interesting: The kids using educational AI tutors aren't becoming robots. They're not losing their humanity or forgetting how to talk to real people. They're just... learning. Research from Harvard's Center for Education Policy Research found that students using DreamBox Learning showed achievement gains of 2.5 percentile points for every 20 minutes of weekly usage, earning the program a 'Strong' evidence rating under ESSA standards. Sometimes better than they would have otherwise. It turns out that having infinite patience and the ability to explain concepts seventeen different ways is actually pretty useful in education. Who knew?

The senators' regulatory impulses reveal something deeper about how we approach new technology when it comes to kids. We oscillate wildly between "this will save education" and "this will destroy our children's souls." There's no middle ground, no nuance, just pure panic or unbridled optimism. It's exhausting, and more importantly, it's not helping anyone.

What if we actually asked kids what they think? The six-year-olds using AI tutors seem pretty happy about it. The teenagers who need help with homework aren't exactly suffering from having access to patient, knowledgeable assistants. Maybe, just maybe, the people we're trying to protect might have some insights about what they actually need protection from—and what they need access to.

The truth is, AI in education isn't going away. It's already here, teaching kids to read, helping with math homework, and yes, occasionally writing essays about symbolism in The Great Gatsby. We can either figure out how to use it responsibly, teaching kids to be critical consumers of AI-generated content while also protecting them from genuinely harmful applications, or we can ban everything and watch as they find workarounds anyway. (Spoiler alert number two: They will absolutely find workarounds. They're teenagers. Finding workarounds is basically their job.)

The distinction between KOSMA's social media restrictions, the GUARD Act's companion chatbot ban, and the embrace of educational AI actually shows something encouraging: legislators learning to draw lines instead of building walls. They're recognizing that "AI for kids" isn't a monolithic category—that the chatbot pretending to be your best friend is fundamentally different from the one teaching you fractions.

The senators and the AI tutors represent two visions of childhood: one where we protect kids by identifying genuine risks and addressing them specifically, and another where we prepare them by giving them tools that actually work. For once, these aren't mutually exclusive. We can protect kids from AI companions designed to manipulate their emotions while simultaneously giving them access to AI tutors designed to help them learn algebra.

Maybe instead of "senators versus AI tutors," we're actually watching the awkward but necessary process of figuring out which AI applications help kids and which ones hurt them. It's not sexy. It's not simple. But it might actually be... progress?

That would require nuance, though. And where's the fun in that?