The U.S. must replace its outdated, test-driven school model with a modern, competency-based platform focused on fundamentals and trainability to prepare a workforce for the AI era.

“As global competition intensifies, the United States must confront a harsh reality of developing a new platform for education as current test scores reflect failure in schools.” – James Carlini, 2025

If the United States hopes to lead the AI-driven economy, it must first fix the machinery that feeds it: public education. The current system—built for another century and another labor market—still measures success with tests that increasingly reveal failure. Emerging technologies such as artificial intelligence, robotics, drones, and 3-D reality demand a highly skilled workforce. We will not meet that demand with today’s outcomes for twelfth graders.

Across the country, many districts are missing the mark on the basics: math, science, and reading. The National Assessment of Educational Progress has released new results, and they are sobering. Scores measured against those from five years ago are down across the board. By the end of 2024, every category had slipped from 2019 levels. Only 22 percent of twelfth graders reached or exceeded NAEP Proficient in mathematics; in reading, just 35 percent did so.

Put bluntly, fewer than one in four high-school seniors are performing at the level they should in math, and nearly two-thirds are not reading at a senior level. That is unacceptable. Students who leave high school without these competencies are unprepared for college, trade programs, or the kinds of advanced training that lead to jobs paying more than minimum wage. Where do they end up? Nowhere good—especially as the jobs most open to them are the first to be automated.

Absenteeism compounds the problem. Students must first show up to learn. In the latest survey, 31 percent of twelfth graders reported missing three or more days of school in the month before taking the assessment in 2024, up from 26 percent in 2019. Should punctuality and attendance be taught explicitly—folded into lesson plans—because too many students aren’t learning them elsewhere? It’s an uncomfortable question, but avoiding it will not improve outcomes.

Parents have a role to play that cannot be outsourced. Children who do not learn more will not earn more, and the market will be unforgiving. Minimum-skill jobs are thinning out as automation takes hold in sector after sector. Robots and robotic systems are poised to replace millions of low-skill roles over the next decade, if not sooner. As the economy generates new work in new industries, students must be prepared early to thrive in workplaces that are growing more complex by the month. Public schools, operated as they are today, are anachronisms—excellent at compliance, poor at capability-building.

Unless we overhaul an institution designed to supply the Industrial Age—nineteenth-century factories, not twenty-first-century labs and data centers—we will continue to fall behind nations that have matched and, in some cases, surpassed us in developing the cognitive competencies today’s economy requires. The United States needs a contemporary platform for education: a coherent, end-to-end framework that educates and equips a qualified workforce for emerging technological fields and for the many domains they will reshape for decades to come.

That platform must begin with first principles and end with employability. It should replace an outdated public-school model—originally meant to transition an agrarian society into an industrial one—with a modern framework that deliberately builds twenty-first-century skill sets for the Age of AI and robotics. Higher education cannot be exempt. As I have said before, we cannot keep teaching checkers when the world has moved on to chess. A modest tweak won’t cut it. What’s needed is a radical shift in competencies and skill sets.

Credentials alone are not the answer. Over-degreed professionals who under-deliver practical results should not be the benchmark. The premium should be on people who can fast-adapt—who are trainable in new domains and comfortable acquiring and applying fresh skills as technologies evolve. Hundreds of CEOs have argued for making computer science and AI education mandatory for every U.S. student if we intend to remain globally competitive. They are right to focus on the pipeline, but the mindset matters just as much.

Trainability should be treated as a core qualification, not a nice-to-have. In an era when acquiring, adopting, and employing new capabilities will be routine, the only job security anyone can claim is the set of skills they carry. That imperative applies to the current workforce, too. Workers who wish to remain viable will have to accept new responsibilities and continually expand their repertoire. Employers, for their part, should insist on and support continuous learning rather than treating professional development as a perk.

The supporting cast of competencies is equally essential: learnability, comfort interacting across platforms, systems thinking, emotional intelligence, and collaborative fluency. In a work environment defined by constant change, these are not “soft” skills. They are survival skills. They determine whether a graduate can adapt to a new tool, process, or team without losing momentum.

The demands of the AI age will reward those who adapt and learn fastest; they will leave behind those who resist. Trainability and learnability are not buzzwords. They are the habits that allow people to learn, unlearn, and relearn as technologies shift. Rapid technological change equals constant reskilling. That cadence should be reflected in how we design curricula and how we signal mastery—through demonstrations of competence, not merely time spent in a seat.

That is why trainability belongs at the base of the platform for education, along with the fundamentals that have always underwritten success: literacy, numeracy, and socio-emotional competence. These are the prerequisites for any higher-order skill, whether in engineering, finance, logistics, or the arts. Without them, everything else is built on sand. They are also the floor on which more advanced cognitive abilities—analysis, synthesis, and judgment—rest.

Most modern projects—across industries—draw on multiple disciplines. Expecting students to collect a series of siloed degrees (accounting here, psychology there, computer science somewhere else) is neither feasible nor efficient. We need more fast-track programs that are shorter in duration and broader in scope, intentionally designed to braid disciplines together. In other words: less silo, more stack. A new platform for education should reward breadth paired with depth, so graduates can move laterally across functions while still drilling deep when needed.

Building this platform will also require new leadership. Obsolete frameworks persist because obsolete incentives and administrators defend them. If we want students to pursue twenty-first-century careers, we need educational stewards who understand twenty-first-century realities—who will push toward measurable competencies, not just defend familiar routines. Replacing outdated governance with leaders who can align curriculum, assessment, and workforce needs is not a cosmetic change; it is the point.

The path forward is not mysterious. Show up. Master the basics. Develop the habits that enable continuous learning. Then apply them across disciplines, repeatedly. A genuine platform for education would sequence those steps from kindergarten through college and into the workplace—not as a patchwork of disconnected reforms, but as a national project.

America’s next golden age will not be financed by nostalgia. It will be built by students, educators, and employers who accept the terms of this era and prepare accordingly. The alternative is drift. And drift, in a world moving this quickly, is just another name for decline.