American schools face a crisis as reading and math scores plummet, with neuroscientist Jared Cooney Horvath linking the decline to the rise of computer-adaptive testing. Utah's National Assessment of Educational Progress data shows a steady downturn since 2014, coinciding with the mandatory adoption of the Student Assessment of Growth and Excellence (SAGE) test. Horvath argues this isn't isolated—it's part of a global trend correlating increased screen time with cognitive decline, marking Gen Z as the first generation to underperform predecessors on standardized assessments.

The $30 billion edtech investment since 2002, including Maine's pioneering laptop program and Google's Chromebook rollout, aimed to personalize learning. Instead, Horvath contends technology replaced proven pedagogy. He cites B.F. Skinner's 1950s "teaching machine" failures, where students mastered tools but not content—a problem now mirrored in AI's rise. A Brookings report found half of U.S. teens use AI for schoolwork, often to cheat rather than learn, undermining critical thinking.

Historical parallels reveal persistent edtech flaws. Skinner's teaching machine and modern AI both prioritize tool mastery over subject mastery. Horvath stresses friction in learning—struggling through problems builds expertise. He warns that offloading cognitive work to AI creates dependency, not skill development. His solution: teach foundational skills first, then integrate technology as a tool for experts, not novices.

Policy implications loom large. With schools spending billions on devices, Horvath calls for curriculum over pedagogy shifts. "Teach math, literacy, numeracy—then let technology enhance mastery," he argues. As AI literacy courses emerge, the challenge remains: can schools balance innovation with evidence-based learning? The data suggests edtech's current trajectory risks creating a generation ill-equipped for critical thinking.