We Are Training Workers for Yesterday’s Jobs: The Institutional Adaptation Gap Threatening America’s Workforce
By: Jason Wingard, PhD
Consider medical coding. The occupation is not disappearing, but the work is changing. AI systems now suggest billing codes directly from clinical notes, and coders increasingly audit machine output and resolve the cases software gets wrong. Some programs have adapted: Thomas Jefferson University now has students evaluate AI-assisted coding against realistic clinical documentation. The risk is that programs that fail to adapt will graduate students with credentials that match the job title but not the work.
The pattern extends well beyond healthcare. Lightcast’s analysis of U.S. job postings found that roughly a third of the skills employers requested for the average job changed between 2021 and 2024.
In The Great Skills Gap, I wrote about the distance between what employers need and how workers arrive prepared. One important cause of that gap is institutional: education, training, and employment systems often struggle to adapt as quickly as the work itself changes. This is the institutional adaptation gap.
I have worked inside that problem. At Columbia’s School of Professional Studies, we launched master’s programs in enterprise risk management, applied analytics, and human capital management to meet clear market demand. Each required review by the University Senate and the provost’s office to ensure academic quality and financial viability. Those reviews protect rigor and institutional credibility. But when occupational requirements change continuously, approval processes measured in months can leave institutions responding to yesterday’s demand. As I argued in The College Devaluation Crisis, the leadership challenge is not to weaken those standards but to redesign how quickly and how often programs can be revised while preserving them.
Employers face a parallel test. As Chief Learning Officer at Goldman Sachs, I helped lead a learning strategy that supported employees at major career transitions, from analyst and associate through vice president, managing director, and partner. The lesson was not that every employer should replicate that model. It was that workforce development should follow the changing demands placed on people throughout their careers, rather than end at hiring.
Three shifts follow.
First, programs should build adaptable capabilities alongside technical skills: the capacity to learn continuously, exercise judgment, and work alongside AI systems.
Second, institutions should create faster approval pathways for updating existing curricula, without eliminating faculty oversight. Modular design, employer partners with real input on content, and regularly updated skills data can make those revisions routine rather than exceptional. Private capital can help finance the technologies, partnerships, and delivery models that make continuous adaptation possible.
Third, accountability should reward adaptability. Credentials earned and jobs filled matter, but so do career progression and earnings measured over years, not months. What we measure shapes what we fund, and what we fund shapes what institutions build.
The question for workforce leaders is no longer only whether we are preparing enough people for today’s jobs. It is whether we are building institutions capable of preparing people for work that has yet to be fully defined.
The greatest workforce risk may not be that workers become obsolete. It is that the institutions responsible for preparing them become obsolete first.
