The tech talent pipeline is reorganizing itself in real time. Nationwide enrollment in computer and information sciences fell more than 8% at four-year institutions this spring compared with the prior year. Entry-level software developer job postings have dropped alongside it. Meanwhile, psychology majors, biology students, and musicians are filling AI courses and signing up for AI minors.
The numbers behind the shift
In 2021, only five U.S. universities offered AI majors. Today there are at least 74 AI majors and 89 minors at American universities, according to data tracked by Northeastern University’s Center for Inclusive Computing. A dozen more AI majors are scheduled to launch in 2026. The University of Michigan began offering an AI minor in winter 2026, open to students across engineering, music, business, pharmacy, and art. Recent arXiv research mapping more than 350 undergraduate AI programs across 569 institutions that produce 86% of U.S. computer science graduates confirms the scale of the expansion.

What universities are actually doing
Schools are moving faster than their usual timelines. Andrew Arroyo, senior vice provost for academic affairs at Virginia Commonwealth University, noted that course design once took 18 months. “Now we’re doing things really in a matter of months and in some cases weeks,” he told Fortune. VCU also plans to launch an online graduate certificate in applied AI this year.
The policy changes are broad. Purdue University added an AI graduation requirement. Harvard weaves large language model lessons into freshman writing classes alongside discussions of copyright and disinformation. Ohio State mandates AI fluency with practical workshops. Northwestern eased prerequisites so non-majors can enter machine learning courses and is recruiting three additional professors to handle demand. The University of Wisconsin-Madison approved a new college merging data science, statistics, information science, and AI, with operations set to launch in July 2026.
Vibe coding and the skills debate
Part of what is pulling non-majors in is the rise of vibe coding: prompting AI to generate functional software without deep technical knowledge. A community college outside Kansas City, Johnson County Community College, already offers a non-credit class on it. Grant Carlson, program coordinator for workforce development there, described the dynamic plainly: “What AI does is shrink that distance between a complete novice and an expert.”
The concern from practitioners is just as plain. Murtaza Ali, a doctoral candidate at the University of Washington who studies computer science education, flagged the risk: “When you ask the AI to do it for you, on the surface it might just seem it is writing the code. But under the hood, it’s actually also doing the understanding of the task for you.” Posts on X note that enterprises now have graduates who cannot write a basic Python script without AI assistance. Security, observability, and compliance still require someone who understands what the code is doing. Leaders at Anthropic reportedly write most of their own code with AI assistance, which cuts both ways in this debate.
The operator takeaway
If you hire developers or work alongside them, the pipeline is changing. The next generation of people who call themselves technical will have a wider range of actual depth. Some will be genuinely capable. Others will be fluent at prompting but shallow on fundamentals. The hiring bar for roles that touch production systems will need to be clearer about which skills actually matter.
