The gap between what universities teach and what the job market demands is widening at a speed no committee can match. AI has already replaced the skills many degree programs certify, and the half-life of a graduate's capability is now measured in years, not decades.
More than a third of entry-level jobs now require AI skills, according to the National Association of Colleges and Employers. That figure has nearly tripled since the fall of 2025. It did not take five years. It took two semesters, one hiring cycle, the length of a single academic year.
A university that reviews curriculum on a seven-year cycle cannot answer a demand curve that moves that fast through a committee. The institution's own timelines work against it. A new degree program can take up to 18 months to move from proposal to approval. A tenure-track faculty search runs 9 to 12 months from posting to a start date. Those processes were built for a world where the job a graduate walked into looked the same at the end of the process as at the start. That world is gone.
Who actually closes the gap
Policy moves at the speed the institution already moves. Only people can close the distance - the ones willing to build inside the cycle while the cycle is still running its slow vote. These individuals exist on every campus. They are mid-level faculty or staff members who found a real problem and built a working answer on borrowed time. The org chart does not account for them, the budget has no line for their work, and the institution's ability to change depends on them anyway.
Gifford Pinchot III coined the word "intrapreneur" in a 1978 white paper and called them "dreamers who do." The university version carries more weight now because the distance they absorb is wider. Anyone can be this person: faculty, staff, a department administrator, someone in the library. Most of them already have the talent. What they need is a name for the role, a tool built for the problem in front of them, and an institution willing to give support and then step out of the way.
Speed is not theoretical
A university can move at scale when the conditions are right. The GI Bill put about eight million veterans through education, and by 1947, half of American college students were veterans. UTeach started as one faculty-built route at the University of Texas at Austin in 1997 and now runs at more than 40 universities, years before anyone wrote it into a plan.
The California State University system, the nation's largest public university system with 22 universities and more than 471,000 students, surveyed its community in the fall of 2025. More than 94,000 students, faculty, and staff responded - the largest study of its kind in higher education. Ninety-five percent of respondents had already used an AI tool. Eighty-two percent of students called AI essential to their profession, and the same share worried about their job security.
Nobody assigned that adoption. It happened because individual people decided the wait was more dangerous than building ahead of policy. The pressure AI puts on a university does not get absorbed by a task force. It gets absorbed by a person, or it does not get absorbed at all.
Three moves to make this month
Name the role, then find the people already doing it. Nobody volunteers for something with no name, and nobody keeps building something nobody has celebrated. These builders come from anywhere: faculty, staff, students. Look for the unofficial tool half a department relies on, the pilot that outlived its funding, the student club that solved something the provost's office is still studying. Name them where others can hear it.
Build the channel before you build the lab. Create one shared, visible place, open to faculty, staff, and students alike, where builders say what they are working on before it is finished. The campus that talks to itself catches two departments building the same project in parallel before either runs out of runway. Keep the bar to enter low and give whoever uses it air cover - a named senior sponsor and one sentence you will repeat unchanged when someone objects.
Put money behind experimentation, not just permission. AI has cut the cost of standing up a pilot low enough that a department can fund one out of its own discretionary budget instead of waiting a year for a line item. Name a person who can approve that spend without a committee, and put a decision date on the calendar beside their name. A review date lets everyone postpone without feeling like they said no. A name attached to a date forces a yes or a no.
Why this matters for education professionals
Every graduate walking across the stage this spring is being measured against a bar that moved twice since they declared their major. The institutions that survive this decade will not be the ones with the most AI tools installed. They will be the ones that kept producing people willing to build before they were told to. For educators and administrators, the most practical step is finding those people already inside the building and connecting them to each other - before someone outside the building does it instead. Building real AI for Education capability starts with the faculty and staff who are already experimenting, often without recognition or resources.
Your membership also unlocks:
