I have spent my whole career in education. I started in Teach For America in South Los Angeles, I co-founded a charter school, and I have spent years now inside an edtech company. So when I say the backlash against classroom technology is mostly right, that is not a view from outside it.
The reckoning is bipartisan, which is the part worth paying attention to. A neuroscientist's Senate testimony went viral this year. Six states enacted screen-time restrictions. Los Angeles Unified voted unanimously to ban district devices through first grade. People who agree on almost nothing agreed on this, and that rarely happens by accident. It happens when an industry has it coming.
A lot of edtech does. The neuroscientist behind that testimony puts the bill at roughly thirty billion dollars of devices pushed into schools, much of it undifferentiated screen time dressed up as progress, and on the undifferentiated part he is right. I am not interested in defending the category. If a child is on a device mainly because the device was there, that is indefensible.
The line the backlash will not draw
What interests me is the distinction the backlash is too blunt to make on its own. Screen time versus no screen time is the wrong fight. What counts is what the student is doing while they are on the screen.
There is a real difference between a child consuming content and a child producing a skill. Watching a video, clicking through a sequence, dragging answers around: the device is doing the work and the student is along for the ride. That is the version worth worrying about, and it is most of what got bought. But a student reading aloud while a tool measures that performance and adjusts in real time is doing something else entirely. They are carrying the cognitive load rather than handing it off.
Here is what I find telling. The same evidence critics use against edtech tends to vindicate the narrow slice of it that has real proof behind it. The neuroscientist everyone is quoting leans on Hattie's 0.40 effect-size threshold to dismiss general classroom technology. His own meta-analysis puts intelligent tutoring systems at a raw effect size around 0.52, comfortably over the bar he set himself. His critique was aimed at undifferentiated screens, not at purpose-built tools with evidence behind them. As he testified: "This is not a debate about rejecting technology. It is a question of aligning educational tools with how human learning actually works." I will take that standard.
What proof has to mean
A screen-time cap is a crude instrument. The reason it feels necessary anyway is that most of this industry cannot answer one question: can you prove this works? A small part of the industry has started staking its own revenue on the answer, which is the only version of that promise I find convincing.
A blanket cap cannot tell the difference between a student watching YouTube on a Chromebook and a student reading aloud to an AI tutor that is tracking their fluency minute by minute, so it takes both. The kids who lose most in that trade are the ones with nobody waiting at home to read with them.
Which is why I would rather argue about proof than about minutes. If a tool cannot show that a child learned more in those ten minutes than they would have any other way, the backlash should take it, and I will not defend it. If it can show that, then we are arguing about the wrong thing entirely. Every child deserves the chance to become a reader, and the least we owe them is the discipline to tell those two cases apart.