In the last twelve months, more than two dozen American universities have launched, announced, or are actively staffing new degree programs built around artificial intelligence. USC, Columbia, Pace, New Mexico State, the University of South Florida, the University at Buffalo, Carnegie Mellon, MIT, Stanford, Northeastern, and dozens of others have new tracks, majors, or entire colleges that did not exist three years ago. The pace is unusual. The framing is even more unusual.

For a junior parent reading the press releases, this looks like good news. Schools are responding. The system is adapting. The right move is to find the program with the strongest AI brand and apply.

That read is incomplete, and the gap between what the press release says and what the program actually is matters more than parents realize.

What’s actually being built

There are four distinct categories of new program, and they’re often described in identical marketing language even though they are very different products.

The first category is the genuinely new academic discipline. MIT’s “AI and decision-making” major sits inside the EECS department and combines computer science, probability, statistics, and decision theory into a coherent four-year curriculum. Carnegie Mellon’s undergraduate AI major does similar work. These programs were designed from scratch by faculty who had to argue internally for the new track, hire for it, and accept that some students who would otherwise have done CS will choose this instead. The curriculum is structurally different from a CS degree. The faculty are different. The career outcomes will be different.

The second category is the cross-disciplinary college or institute. The University at Buffalo’s new AI and Society department, the University of South Florida’s AI and cybersecurity college that enrolled three thousand students in its first semester, USC’s AI school. These are organizational creations that pull faculty from existing departments — engineering, computer science, philosophy, business, sometimes the humanities — and recombine them under a new name. The educational substance varies enormously. Some are real. Some are administrative repackaging.

The third category is the existing degree with a new name. A “Data Science and AI” major that is, on inspection, the same data science curriculum the school has offered since 2018 with two new electives. A “Machine Learning Engineering” track that is the existing CS major plus a required ML course. These programs exist because admissions offices have noticed that students will pay tuition for the right name on the diploma. There is nothing wrong with these programs intrinsically. The substance is real. But they are not new programs.

The fourth category is the marketing announcement that is not yet a program. The university press release says the AI major will launch in fall 2027. The faculty have not been hired. The curriculum is in committee. The first cohort will be enrolled before anyone has taught the upper-division courses. Students who apply to these programs are buying a promise, not a track record.

How to tell which is which

Three questions worth asking, before letting your student fall in love with a program because of the name on the page.

One. How many faculty are listed in the program’s directory, and where did they come from? A real new AI program has ten to thirty faculty members whose research is listed on the program’s site. Their names, their papers, their courses. If the directory is a bulleted list of “affiliated faculty” linking out to other departments, the program is administrative repackaging. That’s not necessarily bad, but it tells you what the student will actually be enrolled in.

Two. What is the required curriculum? A real AI program has a four-year sequence with required courses that build on each other — linear algebra, probability, machine learning, deep learning, a capstone. If the curriculum is “any four of these twelve electives,” the student is getting CS with a different name. Look at the actual course catalog, not the program description.

Three. What happens to graduates? If the program is two or three years old and has graduates, where did they go? If the program is brand new, what’s the school’s track record placing students in adjacent fields? Carnegie Mellon’s undergraduate AI students go to graduate school and to AI labs. A program at a school whose CS placement is weak will not produce different placement just because the major has a different name.

What this means for application strategy

The strategic move is to separate the program from the school’s overall reputation. Most parents work backwards from school name to major. The students who do best in 2026 work the other direction. They figure out what the student wants to study, evaluate which programs can actually deliver that, and then make peace with the school list that follows.

For a student who genuinely wants to study AI, that probably means a top CS program with a strong ML faculty — Berkeley, Carnegie Mellon, MIT, Stanford, Cornell, Michigan, Georgia Tech, Toronto — and an honest read on whether the new AI-branded programs at less-resourced schools will deliver something equivalent. Often they will. Sometimes they won’t.

For a student who is interested in AI as a path to interesting work but doesn’t necessarily want to do the math, the answer is often a strong applied math, statistics, or cognitive science program at a university with active AI research, rather than the AI major at a school with a less developed research base.

For a student who is interested in the social and policy dimensions of AI — how it shapes work, regulation, access, equity — the new AI and Society programs at Buffalo and elsewhere are doing real work. These are interesting destinations and they admit at higher rates than CS or AI engineering majors at the same schools.

The substance, not the slogan

The schools building real AI programs are doing important work, and the students who get into them will benefit. The schools rebranding existing programs are not doing anything wrong, but the rebrand should not be read as a curricular signal. The schools announcing programs that do not yet exist are taking a real risk with the first few cohorts of students who enroll.

The right question for a junior right now is not “which schools have an AI program?” Most do, or will. The right question is “which AI programs have substance behind them, and does the substance match what my student actually wants to learn?”

Those answers are not in the press release. They’re in the faculty directory, the course catalog, and the placement data. Anyone helping you evaluate a program should be looking at all three.

If you’re putting together a college list for a junior with a real interest in AI or an adjacent field, that’s a conversation worth having with the substance in front of you. Press releases are easy to find. The harder questions take a real read.