I spent twenty-seven years at Microsoft building tools that captured data and organized information: spreadsheets, word processors, databases that made life easier at work and at home. Each generation of software made it easier to record, retrieve, and share knowledge. But the thinking itself, analysis, judgment, creative leap, always remained human.

What I didn’t see coming was a technology that captures cognition itself. That is AI. And companies are moving too fast to fully exploit its potential without considering what they’ll lose when humans are taking a backseat to technology in the workplace.

Consider a law firm extracting tremendous value today from using AI to automate tasks like research, drafting briefs, and flagging risks. The trade-off for that short-term satisfaction is the externalization of a thinking process that used to live inside a junior associate’s head. The consequence is depriving that young person of meaningful learning experiences that equip them to become the leaders of the future. Virtually every knowledge profession is vulnerable to this dynamic as companies change how work gets done and, more importantly, where cognition lives.

A society in which fewer people develop the capacity for independent, critical thought is not just less competitive. It is more vulnerable to manipulation, to misinformation, and to the erosion of the informed citizenship that democracy depends on.

For business leaders, this is the defining talent challenge of the decade.

Data confirms the speed of the trend to replace entry-level workers. According to researchers at the Stanford Digital Economy Lab, entry-level workers in AI-exposed occupations have seen employment fall roughly 13% on a relative basis since late 2022, while more experienced workers held steady or grew. And a recent KPMG survey found that more than than half of business leaders expect to reshape entry-level recruiting within the year.

Companies that want to survive this early stage of AI transformation should slow down and embrace the cognitive apprenticeship model through which people build judgment, pattern recognition, and professional instinct before they have a talent debt coming due.

Competency and literacy are not the same thing.

There is no doubt the workforce of the future, and today, will need to effectively manage AI. The companies that will emerge with the most competitive edge in their workforce will be those that understand the difference between competency and literacy in AI.

AI competency is a checklist: prompting tools, summarizing documents, and running analyses in generative platforms. These skills matter, but they are table stakes in this era. AI literacy is more complex and far more valuable.

People who are AI literate use the technology to take their own thinking further, not to replace it. They use AI with enough subject-matter depth to ask it sharper questions, interrogate the logic behind its outputs, and recognize where its analysis may be flawed or biased. Workers with genuine AI literacy know how to use AI as a Socratic partner that challenges their thinking rather than validates it. They generate new ideas, catch expensive mistakes, and lead teams through problems that have no obvious answers. Unfortunately, not enough is being done to ensure the talent pipeline of the future can leverage AI in this way.

The opportunity hiding in plain sight.

That is why business leaders need to treat AI literacy as a national priority. The United States is falling behind. China is mandating AI instruction in all primary and secondary schools. Singapore is training every teacher in AI by 2026. South Korea has rolled out AI-powered curricula across all grade levels, backed by over $800 million in investment. The United Kingdom is investing in national AI educational resources from the primary level up. These are national strategies built on a recognition the United States has been slow to absorb: AI literacy must be developed over years, not acquired in a corporate onboarding session. The EDSAFE AI Alliance’s national Blueprint for Action makes the same case here at home.

The Economy of the Future Commission Act, introduced in March by Senators Warner and Rounds, is a welcome signal. But a commission that reports in a year is not a plan for institutions that need to move now.

The talent is out there. Community colleges alone enroll roughly 41% of all undergraduates in the United States. Add HBCUs and regional state schools, and you are talking about the institutions that educate the vast majority of the future American workforce, institutions largely left out of the national conversation on AI readiness. First-generation students and young people from working families are exactly the people American business needs in its pipeline. Any strategy that overlooks them is leaving competitive advantage on the table.

Brookings Senior Fellow Molly Kinder, whose research focuses on AI’s impact on the workforce, has proposed adapting the medical residency model for white-collar career pathways: structured, mentored programs where learning and doing are the same thing, and where building expertise is the job, not a byproduct of it. Business leaders should champion that model. But it only works if the educational foundation beneath it is strong. A doctor who lacks content knowledge of the community they serve, who cannot communicate across difference, or who cannot recognize the forces shaping a patient’s health, makes worse decisions regardless of residency design. The same is true across every knowledge profession. That foundation must be built across all K-12 classrooms and higher education institutions.

The companies and countries that develop workers who can think alongside AI, not just operate it, will have a durable advantage. If AI is capturing cognition, the most valuable investment American business can make is in the human talent that directs it. The opportunity to build that pipeline, broadly and boldly, is right in front of us.

The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.