Is an AI Major Worth It? Only 141 Students Graduated With One Last Year
The claim: artificial intelligence is remaking higher education, and students who want a piece of it should get a credential that says so.
The evidence: in the most recent federal completions data, 141 students in the entire United States finished an undergraduate degree coded as artificial intelligence. Computer science produced 50,272 in the same year.
Both of those things are true at once. Sorting out what they mean for a tuition check is the whole point of this piece.
What set this off
On August 3, the Associated Press reported that colleges are racing to push AI into every corner of the curriculum. Virginia Commonwealth University stood up an AI minor aimed at students outside computer science. Purdue added an AI graduation requirement, Ohio State an AI fluency requirement. Northwestern launched a minor and a new major. Harvard is teaching large language models inside freshman writing. Johnson County Community College built an AI certificate for workforce students.
Meanwhile the traditional major is shrinking. The National Student Clearinghouse Research Center reported that computer and information sciences enrollment “declined sharply compared to the spring of 2025, down 8.4 percent at four-year institutions and -11.2 percent at two-year institutions.”
So: enthusiasm up, enrollment down. That is a strange market, and strange markets are where students overpay.
How many people actually graduate with an AI degree?
Federal completions data is the cleanest answer available, because it counts degrees that were actually conferred rather than programs that were announced. Every college that receives federal aid reports its completions to IPEDS by six-digit program code.
Here is the 2024 award year, undergraduate only — associate and bachelor’s degrees, first majors:
| Undergraduate major (IPEDS code) | Degrees conferred | Colleges conferring |
|---|---|---|
| Computer Science (11.0701) | 50,272 | 1,024 |
| Computer & Information Sciences, General (11.0101) | 34,556 | 913 |
| Information Technology (11.0103) | 20,296 | 644 |
| Cybersecurity (11.1003) | 17,632 | 640 |
| Artificial Intelligence (11.0102) | 141 | 34 |
Across all 33 computing program codes, American colleges conferred 163,480 undergraduate degrees in 2024. AI accounted for less than one-tenth of one percent of them — roughly one graduate in 1,160.
Two limits on that number, because they matter. It counts associate and bachelor’s degrees only, so the AI master’s programs at places like Carnegie Mellon are not in it. And it counts degrees coded 11.0102 — a program marketed as “AI” can be filed under computer science or under the general computing code with no separate line in the data. The 141 is the floor, not the ceiling.
That caveat cuts both ways, though, and it is the real finding: the credential is so new that the federal classification system has not caught up to it. When a field has no stable code, it also has no comparable outcome data. Hold that thought.
Where do those 141 degrees come from?
Almost entirely from a handful of schools.
- Carnegie Mellon University — 45
- Massachusetts Institute of Technology — 19
- Full Sail University — 13
- Houston Community College — 12 (associate)
- University of Advancing Technology — 10
That is 99 of 141 degrees, or 70%, from five institutions. Carnegie Mellon alone conferred 32% of every undergraduate AI degree in the country.
The tail is thinner than it looks. Of the 34 colleges that reported an AI program in 2024, 11 conferred zero undergraduate degrees — Penn State, Miami Dade College, Ferris State, UMass Dartmouth, Capitol Technology University and others had the program on the books with no associate or bachelor’s graduates that year.
That is not a knock on those schools. It is what a brand-new program looks like in year one or two: approved, catalogued, enrolling, and not yet graduating anyone. It does mean that if you enroll in one, you may be in the first cohort — with no alumni network, no placement history, and no graduate outcome data for the program you are paying for.
Isn’t it growing, though?
Yes, and quickly. AI completions went from 117 in 2023 to 141 in 2024, a 21% jump, and the number of colleges conferring rose from 23 to 34.
Steelmanning the case honestly: those are the growth rates of a credential that is about to become normal, and 11 programs with zero graduates are 11 programs about to produce their first. Anyone reading 141 as “this will never be a real major” is reading it wrong.
But growth rate is not scale, and you are not enrolling in a growth rate. Compare what else grew in the same year:
- Cybersecurity: 13,460 → 17,632 degrees (+31%), now at 640 colleges
- Data science: 1,300 → 2,068 degrees (+59%), now at 205 colleges
- Artificial intelligence: 117 → 141 degrees (+21%), now at 34 colleges
Data science is growing faster than AI and is already fifteen times larger. Cybersecurity is growing at a similar clip from a base 125 times larger. If the goal is a technical degree pointed at the same labor market, two credentials are further along the same road.
Then why did CS degrees go up while enrollment fell?
Because the two numbers are measuring different years, and this is the single easiest way to get fooled reading coverage of this story.
Computer science completions rose from 46,766 in 2023 to 50,272 in 2024 — up 7.5% — during the same stretch that enrollment coverage turned gloomy. There is no contradiction. A bachelor’s degree conferred in the 2023–24 award year belongs to someone who enrolled around 2019 or 2020, at the peak of the CS boom. Completions are a rear-view mirror with roughly a four-year delay.
The spring 2026 enrollment drop will not show up in completions until close to the end of the decade. So when you see “CS degrees are at record highs” and “CS enrollment is collapsing” in the same week, both are accurate, and the second one is the news.
The same lag is why AI’s 141 tells you about 2020-era program offerings, not about what is being built right now.
What does an AI major earn?
Nobody can tell you, and you should be suspicious of anyone who does.
The Department of Education’s College Scorecard publishes field-of-study earnings at the four-digit program level, not the six-digit one. Artificial intelligence (11.0102) sits inside the 11.01 family alongside Computer & Information Sciences General, Information Technology, Informatics, Human-Centered Technology Design and one further code. All six share a single earnings figure: about $82,900 across reporting institutions, or $90,100 for bachelor’s holders specifically.
That figure is real, but it is a six-major blend. It is not AI graduates’ earnings, and with 141 graduates a year nationally there is no cohort large enough to produce one.
Computer science is the useful contrast. It is the only program in the 11.07 family, so its Scorecard number is genuinely CS-specific: about $105,900 for bachelor’s holders, averaged across reporting institutions. That is a figure drawn from a field large enough to produce a stable one. See the full breakdown on our computer science worth-it analysis.
The practical version: for computer science, you can check what graduates earn before you enroll. For artificial intelligence, you cannot. That is a genuine difference in the quality of information you get to make a $100,000 decision with.
So what should you actually do?
- Ask the admissions office one question: how many students graduated from this AI program last year? For roughly a third of the colleges offering one in 2024, the honest answer was zero. That is a fair thing to enroll in with open eyes; it is not a fair thing to discover in year three.
- Ask which program code the degree is filed under. If it is 11.0701, you are getting a computer science degree with AI coursework — which is often the better product, and which employers and graduate programs read more easily.
- Compare the minor against relocating. If the AI major you want is 2,000 miles away and the AI minor at your state school covers the same coursework, the minor wins on cost before the comparison even starts. Most of the programs in the AP story are minors, certificates and requirements — not majors — for exactly this reason.
- Look at the AI-adjacent degrees that already exist near you. Cybersecurity is at 640 colleges and data science at 205. Our data analytics and information technology breakdowns cover what those pay and what they cost, and the full field is mapped in our computing majors hub.
- Consider starting at a community college. Houston Community College, Chandler-Gilbert and Estrella Mountain all conferred AI associate degrees in 2024. Two years of introductory computing at community-college prices, then transfer, keeps your options open while the four-year programs mature — see our transfer pathways data.
- If you want the four-year technical degree, weigh the credential level first. Our computer science path comparison runs associate versus bachelor’s on cost, earnings and time.
The strongest argument against everything above
Peter Stone, who chairs computer science at the University of Texas at Austin, built an AI essentials course for non-majors and told the AP: “In the same way that everybody needs some degree of math, reading and writing, I think everybody needs a degree of AI literacy.” Grant Carlson, who runs workforce development programs at Johnson County Community College, put the mechanism plainly: “What AI does is shrink that distance between a complete novice and an expert.”
If they are right — and the evidence so far suggests they are — then AI competence is heading toward the status of spreadsheet competence: expected everywhere, credentialed almost nowhere. Notice that this argument works against the specialist major, not for it. A skill that everyone needs is taught in a course, a minor, or a general education requirement. That is exactly the shape the AP found colleges building.
The 141 may be low because the major is too new. It may also be low because, for most students, the major is not the right container for the skill.
The verdict
Worth it if: you are admitted to one of the handful of programs with a real graduating cohort and an established research faculty, you want the specialist track rather than general software work, and you have checked the program’s actual placement history rather than its brochure.
Skip it if: the program has not graduated anyone yet, it would cost you more than a comparable computer science degree, or it would require moving away from a school where you could take the same coursework as a minor. In those cases the computer science, cybersecurity or data science degree with AI coursework layered on top gives you the same skills, a labor market that knows how to read your transcript, and — not a small thing — published earnings data you can check before you sign the loan.
The AI boom in higher education is real. The AI degree, so far, is 141 people.
Sources
- Heather Hollingsworth — Associated Press, syndicated by Fortune — “At colleges, the AI boom means everyone wants to dabble in computer science” (Fortune headline: “The computer science major is dying so that the AI literacy minor can be born”) — August 3, 2026 — https://fortune.com/2026/08/03/vibe-coding-college-ai-minors-computer-science-decline/
- National Student Clearinghouse Research Center — “Final Spring Enrollment Trends,” spring 2026 — June 4, 2026 — https://nscresearchcenter.org/final-spring-enrollment-trends/
- National Center for Education Statistics — IPEDS Completions, 2024 award year (first majors, associate and bachelor’s degrees), analyzed in-house — https://nces.ed.gov/ipeds/
- U.S. Department of Education — College Scorecard, field-of-study earnings (four-digit CIP, undergraduate credential levels) — https://collegescorecard.ed.gov/data/
