ISSUE 012
Who Teaches AI at the Two-Year Level
I wanted a straight answer to who teaches applied AI at the two-year level, and to whom. What I found is that the people best positioned to count have not counted, that the largest federal investment in building that capacity is a single grant, and that a good deal of the curriculum showing up in catalogues is donated by a software vendor. That is not a complaint. It is the shape of a field that is about three years old.
An idea worth thinking about
The most comprehensive map of American AI programmes stops at the four-year line
In May 2026 a team led by Felix Muzny and Carla Brodley published the most thorough census yet of undergraduate AI programmes in the United States. They built scraping and mapping tools, searched more than 560 institutions covering 86 percent of U.S. computer science graduates, and identified more than 350 AI programmes — majors, minors, concentrations and certificates. They then analysed the course requirements of 66 majors and 87 minors in detail.
The scope line sits in the description of the dataset: four-year universities. Two-year colleges are outside it. That is a defensible research choice and the authors do not hide it. It also means the figure most often quoted for how many AI programmes exist in America systematically excludes the institutions that enrol the largest share of American postsecondary students by headcount.
The authors are careful about their own limits. This is a spring 2026 snapshot, the tool updates as programmes appear, the sample is defined by which institutions were searched rather than by a complete census, and the course analysis rests on a subset of what they found. Their claim to be the most comprehensive snapshot to date is their own assessment, and the paper is a preprint rather than peer-reviewed work. I find it credible, which is exactly why the gap at the bottom of the frame is worth naming.
A tool I disappeared into
The federal database where you can check your own local college, one classification code at a time
The Integrated Postsecondary Education Data System is where every institution receiving federal student aid reports, and it is open to anybody. The Trend Generator shows movement over time in enrolment, completions, graduation rates and finances. Compare Institutions downloads data for more than 7,000 institutions with up to 250 variables as CSV. Summary Tables builds custom tables for a subgroup of institutions, including degrees and certificates awarded.
The key that makes it usable is the classification code. Artificial intelligence has its own CIP code, 11.0102, which is how a two-semester technical certificate at a small college becomes findable in the same dataset as a doctoral programme. If you want to know whether the community college nearest you is awarding AI credentials and to how many people, this is where that answer lives.
Two honest limits before you spend an evening in it. IPEDS completions run roughly two years behind, so a programme that launched last autumn will not appear for a while. And non-credit workforce training, which is how most community colleges actually deliver short AI courses to working adults, is largely invisible in completions data. The NCES landing page also does not spell out which filters each tool supports, so plan on some clicking.
Learn something
Read one actual two-year AI certificate, course by course
Marion Technical College, a two-year technical college in Ohio, publishes its 2026-2027 catalogue online, and its Technical Certificate in Artificial Intelligence, filed under CIP code 11.0102, runs 31 credit hours across two semesters. I am linking it not because Ohio matters to you but because it is the clearest published example I found of what this credential actually contains.
Autumn, 16 credits: Introduction to Information Technology, Intro to Programming Concepts with Python, Introduction to Artificial Intelligence, English Composition I, Academic and Career Success, Applied Mathematics for AI. Spring, 15 credits: Web Application Development, Intro to Logic and Visual Programming, Machine Learning, Natural Language Processing, Computer Applications.
Notice what that is and what it is not. It is a programming credential with AI coursework layered on: Python first, then machine learning and natural language processing. It is not a course in using AI tools to run a business, which is a different thing that two-year colleges mostly deliver as non-credit workforce training and which shows up in no national dataset. If you are advising somebody about going back to school, the distinction between those two paths is the entire conversation.
What we’re seeing
There are 1,015 community colleges serving 10.6 million students, and the federal programme built to give them AI curriculum began with a single $2.8 million grant
The American Association of Community Colleges publishes a Fast Facts sheet each spring. The 2026 edition counts 1,015 community colleges in its membership database as of January 2026: 923 public, 56 independent, 36 tribal. Autumn 2024 enrolment was 6.5 million for credit, split 2.1 million full-time and 4.4 million part-time, plus 4.1 million non-credit students, for 10.6 million in total. The average student is 27 and the median is 23. Sixty-seven percent attend part-time. Thirty-two percent are the first in their family to go.
In August 2024 the National Science Foundation awarded Miami Dade College $2.8 million through its Advanced Technological Education programme to establish the National Applied Artificial Intelligence Consortium, whose job is to help two-year institutions develop technician-level AI courses and certification programmes and to build a national community of practice for faculty. NSF's announcement names no other institutions, sets no student or faculty targets, and does not say how many colleges are involved.
That is one grant for a sector of a thousand colleges. The arithmetic is mine rather than NSF's, and ATE is not the only funding stream — states, Perkins dollars and local employers all pay for workforce programmes that never appear in a federal press release. But it is a fair indication that the capacity to teach this is being assembled grant by grant rather than built as a system.
Some of the gap is filled by vendors. Intel described an AI for Workforce programme with community colleges on its own corporate blog starting in 2021, and a trade publication reported in April 2026, citing a LinkedIn post by the consortium's founder, that Microsoft is expanding free curriculum, certification bootcamps and software access through NAAIC. I am naming that sourcing chain rather than presenting either as established fact. Donated curriculum is genuinely useful to a college with no AI faculty and no budget to hire one, and it is also a channel for teaching one company's toolchain. Both are true at once, and a prospective student deserves to know which they are enrolling in.
In Missouri the visible movement is at the universities rather than the two-year colleges. The UM System Board of Curators approved master's programmes in artificial intelligence at Mizzou and UMKC and a doctorate at Mizzou in early February 2026, with MU's 15-credit AI certificate having begun that spring, and the new programmes still requiring state approval. I could not find a comparable credential at a Missouri community college. I want to be precise: I could not find it is not the same as it does not exist. IPEDS lags by about two years, non-credit workforce training is not reported at all, and college catalogues are not indexed consistently enough to trust a search.
So, to the question this entry asked. Who is actually teaching this, and to whom. The honest answer is that a small number of two-year colleges are teaching programming-centred AI certificates to a small number of students, a larger number are running short non-credit courses that nobody counts, and the national figures everyone quotes come from four-year institutions. If that is unsatisfying, it should be. It is also a reason to ring the college nearest you and ask directly, because they will tell you, and no dataset will.
One question
Have you ever hired someone out of a two-year college?
I am trying to learn whether the credential matters to employers around here or whether you hire on the person in front of you. Reply with what you have actually done, not what you would say at a panel.
More in this pillar as it's written.
— Shana
Sources link to the original publisher, never to a summary.