America Is Building Super Intelligence. Who Is Training Every Student to Make AI a Superpower?

As the White House advances its SI agenda, the AI Institute argues that training human judgment could create a compounding national advantage.

By: Artificial Intelligence Institute

America is accelerating the development and application of artificial intelligence. At the White House's October 8 science summit, the Trump administration used the language of super intelligence, SI, to describe the technological capabilities it hopes to advance in scientific discovery and national competitiveness.

That ambition is bold, visionary and substantial. It deserves an equally ambitious investment in the people who will use those systems.

The AI Institute has a simple way of describing what is at stake: properly trained, a person's AI can become a superpower. Not because the machine substitutes for their mind, but because it expands what their own mind can explore, question, verify and accomplish.

The missing training layer

Students already have access to systems that explain, summarize and answer questions. AI can unlock extraordinary learning. It can also tempt a learner to delegate the thinking itself. In the Institute's classrooms, 78% of students were already using AI before anyone taught them how, and only 13.6% had any guidance.

The difference is cognitive offloading versus cognitive amplification. In one case, the student asks AI to deliver the result. In the other, the student brings the intention, frames the challenge, examines alternatives, verifies information and remains the author of the outcome.

This is not about restricting a powerful technology. It is about teaching a person to command it.

MACH 5: fundamentals before acceleration

The AI Institute's MACH 5 pathway is organized around five lessons, five hours and five repeatable behaviors:

THINK → FRAME → QUESTION → VERIFY → COMMAND

The design draws on developmentally appropriate behavioral engineering, cognitive scaffolding, guided metacognition, repetition and memory anchors. The aim is to build dependable, transferable habits of intelligent human-AI interaction, what the Institute calls mental muscle memory.

Over two school years across more than 20 school sites, the Institute's own results show students moving from answer-seeking toward questioning, verification and ownership of their thinking. Students using AI the right way rose from 15.8% to 88%. Students verifying what AI tells them rose from 14% to 89%. Students using it safely rose from 14% to 90%. The approach has travelled upward too: a struggling third-year aerospace engineering student used Professor Atlas and the program, reached 95% and 93% in his toughest classes, and stayed in aerospace engineering. These are first-party results. Independent evaluation and longitudinal measurement remain essential, and this is not a claim of permanent neurological change.

From school evidence to national interest

The classroom results have drawn serious attention beyond the districts where the work began. The National Science Foundation has met with the Institute and encouraged an independent university research partnership to study the results. The U.S. Department of Education asked the Institute to prepare an exemplar for the Secretary's toolkit process, and it has been delivered.

The Gates Foundation has shaped American education and global health on a scale few institutions in history can match. The Institute is honored that its classroom results reached the foundation's K-12 team, which requested a meeting, and the two have since held several formal meetings at the foundation's request. A meeting is not an endorsement, and the Institute does not claim one. But it is a clear sign that these field results are beginning to travel.

That progression now reaches Washington. The Institute has been invited to brief White House staff in person at 3:00 p.m. on October 15 to discuss field evidence, teacher preparation and the national human-readiness opportunity. A scheduled staff briefing is not a federal endorsement, funding commitment or approval of a national program.

At the center of the conversation is the Institute's 74-Cent Moonshot: a proposal to make essential, tool-independent AI readiness training available to every K-12 student in America at no cost to school districts. The approximately 74-cent figure expresses the proposed initiative's total ten-year cost spread across the national K-12 student population, roughly the cost of a pencil per student. It is not a per-course price or an annual charge. The aim is to remove the cost barrier that too often keeps foundational training from reaching every community.

The Institute sees complementary paths toward that goal. Philanthropic support could provide the resources to build and sustain broad access; federal and state education leadership could help bring districts, educators and implementation partners together at national scale. Neither path is committed or approved today. But the moment is immediate: America can pair its investments in increasingly powerful intelligence with practical training for the people expected to use it.

Fundamentals travel, from a classroom to the boardroom

Brian Taylor, PhD, a Princeton graduate, two-time professional basketball champion and longtime educational leader, understands the power of repeated fundamentals. An athlete does not gain mastery simply by being handed elite equipment. Mastery comes from practice, disciplined judgment and learning how to respond under pressure.

AI is no different. Its fundamentals matter to an elementary learner, a university researcher, a teacher, a CEO and a family-office principal. Proper training does not give everyone the same answer. It gives more people a better method for finding and evaluating answers themselves.

A workforce advantage that could compound

Imagine a graduating class arriving in higher education and employment already accustomed to questioning AI, checking sources and using it to deepen problem-solving rather than bypass it. A year later, another cohort arrives. Then another. And another.

If training is reinforced across grades and institutions, each successive class could bring organizations better shared norms of inquiry, verification and human command. Over years, those habits could accumulate into a broader improvement in workforce capability and institutional decision quality. This is a strategic projection, not an already measured national productivity result. But it points to an extraordinary possibility: the advantage from better AI training may grow every year that trained learners enter the workforce.

The human is the operating system

Super intelligence may dramatically expand the capabilities of our tools. The Institute's complementary mission is to expand the capabilities of the people holding them. Not simply AI literacy. Not only access. A practical foundation in how to think with AI, reason through complexity, verify claims and remain in command.

The country's greatest AI superpower may be millions of people trained to use it well.

Think First. AI Helps Second.