Universities Are Launching AI Degrees This Fall and Graduates Are Asking What They Buy

Fast Company reported on September 4 that universities are adding bachelor's degrees in artificial intelligence while the students who would take them are increasingly sceptical about what an entry-level job will look like. The scepticism is reasonable, and the numbers behind it are worth setting out properly. At Metaintro, we look at new qualifications the same way an employer eventually will, which is by asking what job it maps onto and how many of those jobs exist. Here is what these degrees are, what the labour market data says, and how to decide whether one is worth four years.

Which Universities Are Actually Launching These Degrees?

A genuine wave rather than one or two outliers, and they are ordinary state and private universities rather than only elite research institutions. The University of Kentucky is accepting applications for a bachelor's degree in artificial intelligence. Kennesaw State University is launching a Bachelor of Science in Artificial Intelligence for autumn 2026 on its Marietta campus and online. UNC Charlotte is opening a BS in AI through its College of Computing and Informatics, and the University of New Haven is introducing one built around technical, ethical and industry-focused skills.

The significance is the institutional spread rather than any single programme. When a specialism moves from being a concentration inside a computer science degree to being its own named credential at a range of universities, it usually reflects student demand and employer marketing more than a change in the underlying body of knowledge. Curricula take years to build and universities compete on programme names, so a wave like this tells you what applicants are asking for.

That timing is worth holding in mind. A degree launching in autumn 2026 produces its first graduates in 2030. Anyone starting one is making a four-year bet on what a labour market will value at the end of a period in which the technology has been changing every few months. That is not an argument against the degree, but it is an argument for choosing one whose foundations outlast any particular tool.

What Does the Graduate Job Market Look Like Right Now?

Difficult, and worse for graduates than for workers generally, which is the reversal of the historical pattern. The Federal Reserve Bank of New York reported that conditions remained challenging for recent college graduates through the second quarter of 2026, with an unemployment rate of around 5.6 percent and an underemployment rate of 42 percent. Fast Company cited a similar comparison of 5.7 percent for recent graduates against 4.1 percent for the overall workforce.

The underemployment figure is the one that should shape decisions. Roughly two in five recent graduates who are working are in jobs that do not require a degree at all. A qualification is only worth its cost if it moves you out of that category, and the evidence that any particular degree reliably does so has become weaker across the board rather than only in technology.

The tasks that historically got graduates through the door are also the ones most exposed. Entry-level work has traditionally meant doing the volume tasks that senior people do not have time for, and those are precisely the tasks the current tools handle. That is the mechanism behind AI coming for entry-level jobs first, and it is why the 2026 hiring rebound has been skipping new graduates even as overall payrolls grew.

What Job Does an AI Degree Actually Lead To?

This is where the data does something useful, because the job most people imagine is much smaller than the discourse suggests. The US Bureau of Labor Statistics tracks computer and information research scientists, the category that covers the research end of artificial intelligence work. Median pay in 2025 was $140,300, and projected growth to 2035 is 22 percent, which the agency classes as much faster than average.

Now read the other two numbers in the same table. The occupation held 38,600 jobs in 2025, and the 22 percent growth translates into roughly 8,400 additional positions over the entire decade. The typical entry requirement is a master's degree. That is an excellent career and a very small door, and a bachelor's degree in AI does not open it directly.

The realistic destination is considerably larger. Software developers, quality assurance analysts and testers numbered 1,905,400 in 2025, with median pay of $134,040, a bachelor's degree as the typical entry requirement, and 10 percent growth adding 185,400 positions to 2035. The pay is within $6,000 of the research role, the field is fifty times larger, and a bachelor's degree is the standard entry point. Anyone weighing an AI degree should be comparing it against a computer science degree that leads into that occupation, not against an imagined research career.

Is a Specialised Degree Better Than a General One?

It depends entirely on how durable the specialism is, and this is the real question underneath the trend. A degree named after a technology carries a risk that a degree named after a discipline does not, because the technology can change faster than the curriculum. Computer science has survived several complete turnovers of dominant tools because it teaches the foundations underneath them.

The strongest AI programmes are effectively computer science degrees with a different emphasis, covering mathematics, statistics, systems and software engineering alongside machine learning. The weakest are collections of tool training that will need rewriting before the first cohort graduates. When comparing programmes, the module list matters far more than the programme name, and the specific question to ask an admissions office is how much mathematics and core computing the degree requires.

There is a counterintuitive finding worth weighing here too. Some of the most valuable roles emerging around this technology are not the most technical ones, which is why the best AI jobs may go to humanities majors in some organisations. Judgment about where a system should and should not be used, how to evaluate output, and how to communicate limitations to non-technical colleagues is scarce, and it is not taught by a machine learning module. Testing whether a major will hold its value against AI turns on exactly that distinction.

Why Are Students Sceptical?

Because they can see the entry-level market they are about to enter, and it is not reassuring. The Fast Company piece, written by a recent graduate, describes a question now appearing routinely in applications and interviews about how candidates integrate AI into their work, and notes the bind it creates. Use it too much and you look lazy. Avoid it entirely and you look out of touch. Nobody has told candidates which answer is wanted.

That uncertainty is the honest state of the market rather than a failure of preparation. Employers are themselves working out what they want from entry-level hires in this environment, which is why the signals candidates receive are contradictory. Some firms are shrinking graduate intakes, others are raising what they expect new joiners to arrive knowing, and both are responses to the same underlying change.

Scepticism about a specific degree is also rational given the broader shift in how people value qualifications. A third of people now say a degree is not worth it, and that view has hardened as graduate underemployment has risen. A student weighing an AI degree in 2026 is not being cynical by asking what it buys. They are asking the question the data suggests they should.

How Should You Decide Whether to Take One?

Judge the curriculum rather than the title. Ask how much mathematics, statistics and core computing the programme requires, because those are what transfer if the tooling changes entirely. A degree that is 70 percent foundations and 30 percent current techniques will still be worth something in 2032. One built the other way round may not be.

Ask what the programme's graduates actually do. New degrees have no alumni outcomes, so ask instead what the department's existing computing graduates do, where they are placed, and which employers recruit on campus. A brand new AI degree inside a department with strong industry links is a very different proposition from the same degree at an institution with none.

Compare the alternative honestly. For most people the realistic comparison is an AI degree against a computer science degree with AI electives, and the second option is usually more flexible, more widely recognised by employers and identical in the parts that matter. There are also non-degree routes into technical work worth weighing, from digital credentials that employers will actually recognise to apprenticeship-style entry, and the calculation has changed enough that they deserve a serious look rather than an automatic dismissal.

What Does the Rest of the World Say About This Bet?

Other education systems are making the same wager, and watching them is instructive because some moved earlier. Chinese universities have been cutting language majors to make room for AI programmes, which is a more aggressive reallocation than anything happening in the US and a useful preview of the consequences. When a system converts capacity from one discipline to another at speed, the cohort caught in the transition carries the risk.

China also offers a caution about volume. Its universities are producing record numbers of graduates into a difficult market, which demonstrates something universally true and frequently forgotten: a degree's value depends on how many other people hold it. If AI degrees prove popular, the graduating cohorts of 2030 and 2031 will be competing against each other for a set of roles that the occupational projections suggest is growing steadily rather than explosively.

The broader lesson is about timing rather than subject. Qualifications that everyone identifies as valuable at the same moment tend to be crowded by the time the first cohorts graduate, while the skills that stay scarce are the ones that are harder to package into a programme name. That is not a reason to avoid a technical degree. It is a reason to make sure the degree teaches something durable underneath the label, because the label itself will be common by the time you use it.

What If You Are Already Partway Through a Degree?

Do not switch on the strength of a programme name. If you are in a computer science, mathematics, statistics or engineering degree, you are already studying the foundations that AI work is built on, and adding electives plus demonstrable project work achieves most of what a rebranded degree would. Employers hiring for technical roles look at what you can do far more than at what your degree was called.

If you are in a non-technical degree, the useful move is not to abandon it. It is to add the specific, evidenced capability that pairs with it. A humanities or social science graduate who can genuinely evaluate model output, understands the limits, and can explain them to a non-technical audience is filling a gap that most technical graduates do not fill. That combination is currently scarce in a way that another computer science graduate is not.

In both cases, build a portfolio rather than collecting certificates. Two or three real projects where you can explain what you built, what went wrong and how you verified the result will outperform a list of course completions in any interview. The market for entry-level technical work is competitive enough that demonstrated work is doing the job a degree name used to do, which is a harder standard but a more accessible one.

Related Articles

• The Best Employers for New Graduates in America
• Arkansas State University Launches New AI Courses
• What You Can Do With an Occupational Studies Degree and Which Jobs Hire
• College Grads Hit a Job Wall
• What Graduates of an Earlier Recession Want Today's Graduates to Know
• Best Jobs for a Career Change and Why Transferable Skills Matter More
• New Aviation Mechanics Are Landing Jobs Before They Graduate
• Job Market Freeze and the College Student Internship Crisis
• Are Coding Bootcamps Worth the Investment
• Skilled Trades Paying Six Figures With No Degree Required
• Why a Wall Street Fund Just Hired New Grads for the First Time

People Also Asked

Q: Is a bachelor's degree in AI worth it?

A: It depends on the curriculum rather than the name. Programmes weighted toward mathematics, statistics and core computing teach foundations that survive changes in tooling, while programmes built around current tools risk dating quickly. The realistic destination for most graduates is software development, which had 1,905,400 jobs and median pay of $134,040 in 2025 and takes a bachelor's degree, rather than AI research, which had 38,600 jobs and typically requires a master's.

Q: How hard is the job market for new graduates right now?

A: Harder than for the workforce overall, which is unusual. The Federal Reserve Bank of New York reported recent graduate unemployment of about 5.6 percent in the second quarter of 2026 against roughly 4.1 percent for all workers, with an underemployment rate of 42 percent, meaning about two in five working graduates are in roles that do not require a degree.

Q: Should I switch my major to AI?

A: Usually not if you are already in computer science, mathematics, statistics or engineering, because those degrees teach the same foundations and are more widely recognised. Adding AI electives and building two or three demonstrable projects achieves most of the benefit without the risk of a programme name that ages badly. If you are in a non-technical degree, adding evidenced capability alongside it is generally stronger than starting over.

Choosing a degree is a four-year bet on a labour market that is changing every few months, and the safest bets are the ones that keep the most doors open. Metaintro tracks what employers are actually asking for in entry-level technical roles, including which qualifications appear in job postings and which do not. Join free to see the requirements before you commit to the course.