We’ve all heard these warnings over the years. Screens are damaging our children’s brains, and it seems to have reached a tipping point for teachers, administrators and parents.
Have test results was suspended and attention covers appear shorter– although I’m always a bit skeptical about how we measure it. As part of the congressional testimony, a researcher positions that Gen Z may be the first generation in modern history to underperform their parents on cognitive assessments. In his testimony, Dr. Jared Cooney Horvath spoke about classrooms full of digital distractions. Today’s kids spend more time with technology than any generation, and somehow our literacy and problem-solving skills are failing us.
Yes, the concern is real. But the conversation focuses on a difference that may be more important than that screen time The screen actually asks the child what it is.
what Sesame Street He already taught us
We’ve done this experiment before, just with a different box.
In the age of television, researchers found that the impact of screen exposure depends less on the hours logged than on what the child was doing during it—context over content. “Screen time” is like Sesame Street produced measurable benefitwhile passive viewing and background television are associated with poorer language development and short attention spans.
The same medium. Different requirements for the viewer. Different result. I want to say that it is very simple and for television it can be.
Where the analogy breaks down
Sesame Street can make a child. It can ask questions, give pause, invite speculation. What it failed to do, a key point of difference, was to respond to the response. The participation it required was real because the show didn’t know what the child said in response. The screen has never been adjusted.
But today, the great model of language is not only adjusted, but it is repeated. And along the way, this unique form of engagement changes the cognitive geometry of the box even further.”student centered” dynamic. And that changes the meaning of “interactive” in a way that I don’t think we’ve fully accounted for.
Used appropriately, LLM can do the following Sesame Streetfaster cousin. A student works with an idea and discovers something new. But it can easily behave like an automaton that only talks back. AI responds in a cool and transactional way. From the whole room, these are the same. Baby, screen, type a little. Only one of them involves thinking, and you can’t always tell which one from the outside.
This is the difference I believe in the screen time debate remains missing. Task switching and burnout are real problems, but they’re not the whole story—perhaps not even the main story. A student going through five fun answer sheets and a student working on a hard question are technically doing the same activity. Both AIs are open in a window. But I don’t think it’s even close to a semantic similarity.
Friction is an educational program
Learning is messy.
A student encounters a new concept, misreads it, revises it, challenges it, and even has to start over. The process is slow and maybe even a little embarrassing. (Perhaps this is what we are trying to eliminate.) But this struggle or struggle is not ineffectiveness in learning. He learns by himself.
AI can support this process. It can also short-circuit it, and short-circuiting is seductive precisely because it clearly sounds and feels authoritative. The danger is that polished output reads like insight, even when no insight has occurred.
So maybe there is no AI line and no AI. This is whether the student provided the “intellectual impetus” before the LLM came along, or whether the model provided it for them.
A rule of thumb, not a policy
I don’t think this should be resolved into a clean classroom protocol, and I’m not the teacher to suggest it. Also, I’m skeptical of anyone who claims this. But here is a rough version. Let the student work on the problem alone for a while first, if necessary. Let them engage with the AI with a realistic position, even a wrong one. So let the conversation complicate what conversations do.
Then ask the student to explain the idea by closing the screen. Ask where the model changed her mind and if she can defend why she should have. For me, it’s the ability to explain your work – from school to business – that’s the key aspect here that reveals the learning process, not just reciting facts.
None of this is feasible at scale, and I suspect most classrooms won’t bother because the school day isn’t built for this slow pace. This is a separate problem from the one I’m describing here, though probably unrelated.
Apart from the screen time number
Parents already feel this difference intuitively, even if they describe it differently. Not all screen activities are the same because they don’t require the same thing from the user.
Educational policy measures the wrong variable. Counting minutes is easy. Not a cognitive requirement, perhaps the real reason we count the minutes.
The more difficult question, which I don’t think Horvath’s testimony, or much of the coverage surrounding it, gets enough of is what happens when a screen stops being a passive thing that a child does and becomes something that makes you think.
Good job, it’s not a threat. It’s repetitive engagement, the same mechanism that made Sesame Street work, except now the conversation never stops. A model who constantly pushes, asks what you mean, doesn’t let a half-baked idea pass as a finished idea, is perhaps the most demanding and powerful mentor a student can have. And this is a very protective version.
But it’s also a version that’s likely to get buried under an easier version. The replay takes longer than the answer. It asks more of the child, not less, and asking for more is never a feature that goes to market. We do not have Sesame Street learning which version actually wins at scale. We are running this experience through our own children and I don’t know which version most of them are getting now.
