The AI Major Nearly Doubled. Your Course Still Needs Its Own AI Outcome.

The AI major nearly doubled this year. It is training people to build AI. Your course is where everyone else learns to judge it.

WBUR reported last week that AI majors housed in university computer science departments have nearly doubled since last December, citing a database kept by Northeastern University’s Center for Inclusive Computing. The center’s live map listed 108 AI majors at 103 schools on October 30, alongside 134 minors, 188 concentrations and 172 certificate style tracks.

That is a real shift, and the programs behind it are serious. It is also the wrong number for most faculty to stare at. Every one of those programs is something a student has to choose. The nursing major in your section did not choose one. Neither did the history major, the accounting major or the student who has not declared anything yet. They will still meet AI drafts in their fields, and for many of them the place they practice judging those drafts will be a course like yours.

What the AI major boom actually tells faculty

Start with what the programs say about themselves, because it is more useful than the headline.

At MIT, the AI and Decision Making major launched in fall 2022 and is now the second most popular major on campus, with 354 enrolled students, according to the registrar’s office as reported by WBUR. Its faculty head, Antonio Torralba, was direct about the goal: “Our goal is not to teach them to be users of [generative] AI, our goal is to teach them the foundations needed to build the future of AI.”

Northeastern renamed its bachelor’s degree in data science to AI this fall. Felix Muzny, a computer science professor there, described the same aim: “We are training people to be engineers of AI systems, so not just able to use them, not just understand how they work, but understand how to go inside and change them.”

At Bentley University, the AI major opened in fall 2024 with a handful of students. This year nearly 100 undergraduates are pursuing it. Bentley’s president, Brent Chrite, named what is driving the demand: “We see the angst, and we see the anxiety.”

Put those three together and the picture is clear. The AI majors are building engineers and builders, and they are growing because students are anxious about an AI economy. That anxiety does not stay inside the computer science building. It sits in every one of your sections.

Even the AI majors do not all require judgment

Here is the detail worth reading twice. In a study the same Northeastern team posted in May, the researchers analyzed the requirements of 66 AI majors and 87 AI minors. More than a third of the majors require an Ethics in AI course. Only 24 percent of the minors do. Read that the other way around: at the rest of those majors, and at roughly three in four of those minors, a student can finish a program whose whole subject is AI without a required ethics course.

Some programs are closing that gap on purpose. Worcester Polytechnic Institute’s new bachelor’s degree in AI requires every student to take a course on responsible AI. The program head, Elke Rundensteiner, put the reason plainly: “The students need to understand when to use [AI] and how to make it reliable.”

When to use it. How to make it reliable. That is not a computer science skill. It is a disciplinary one. Whether an AI summary of a primary source is reliable is a history question. Whether an AI drafted care plan is reliable is a nursing question. Whether an AI reconciliation ties out is an accounting question. The person qualified to teach that judgment in your field is the person who already grades it.

The quote that should change your spring planning

Zack Mabel, interim director of the Georgetown University Center on Education and the Workforce, told WBUR that incorporating AI should extend beyond degree programs: “Curricula of all different flavors need to embed AI as part of what they are teaching and equipping students to be able to do.”

Curricula of all different flavors. That is your course, not a new one. And notice what he did not say. He did not say every course needs a unit about AI, a guest lecture on AI, or a policy paragraph about AI. He said embed it as part of what students are equipped to do.

I read that as a thread, not a subject. AI literacy built as a separate course reaches the students who sign up for it. AI literacy built as one outcome inside every discipline reaches everyone, and it is taught by the person who knows what good work in that discipline looks like.

The AI judgment assignment, in four moves

You do not need a new course, a committee or a vendor to start. You need one assignment. Here is the method, and it works by hand.

Seed: pick the task AI already does for your students. Not the task you wish they would not use it for. The one they already hand to it. A summary, a first draft, a problem set solution, a lab discussion section, a citation list. Be honest about it, because the assignment only works if it meets students where they already are.

Sprout: name the standard your field checks first. When an expert in your discipline looks at that kind of work, what do they check before anything else? The source. The calculation. The protocol. The evidence for the claim. Write it in the words you use when you grade.

Grow: write the outcome and the assignment. The outcome takes one sentence: students will judge AI output for that task against that standard and correct what fails, with evidence. The assignment follows from it. Students get an AI draft of the task. They mark every place it fails the standard. For each mark they show the evidence, from the reading, the data or the method. Then they hand in the corrected version alongside the marked one.

Harvest: write one rubric row. Three levels are enough. At the top, the student found the failures that matter and proved each one. In the middle, the student found some and asserted more than they proved. At the bottom, the student polished the AI draft without testing it. That last level is the one you are really trying to grade out of existence.

That is one outcome, one assignment and one rubric row. It fits inside the course you already teach.

Why this beats banning it or ignoring it

A ban tests whether students can avoid a tool they will use at work. Ignoring it means the AI draft gets graded as if the student wrote it. The judgment assignment does something neither can: it makes the AI output the object of study, and it grades the thing that will still matter after the tools change, which is whether a person in your field can tell good work from plausible work.

It also answers the anxiety Bentley’s president described with something better than a new major. Rundensteiner told WBUR she hears from advisory board members representing companies like Microsoft that they cannot find enough people skilled in the latest AI technologies. Whether that holds in every field is an open question. What does hold is that a graduate who can check AI output against a professional standard is useful in any of them.

Mabel framed the goal as preparing students “with the lasting skills and the ability to learn so that they can adapt to changes in the future.” A tool changes every few months. The standard your field checks first does not.

Start with one assignment this spring

The free session walks you through the four moves on your own course. You will leave holding a finished assignment and a rubric row, written against your field’s standard, ready to drop into your spring course.

Then the next thought arrives on its own. That is one assignment for one course. You probably teach more than one, each with its own standard, its own tasks and its own rubrics to write. The Assessment Builder builds assignments, rubrics and tagged question banks, grounded in your Course Record, the record of your course, your outcomes, your policies and your voice that every tool in the toolkit uses. Full access for 14 days, cancel in one click. Build the first judgment assignment by hand anyway, so you know exactly what a good one has to hold.

If you would rather work it out with other instructors first, the Harvest Kernel Learning Community is free, and this is the question there this week.

The AI majors will train the people who build the tools. Your course decides whether everyone else can tell when the tools are wrong.

Sources

• Suevon Lee, “In higher education, the race is on to offer AI as an undergraduate major,” WBUR, October 23, 2026. wbur.org
• “Why Northeastern created a database to track AI majors at universities,” WBUR Today newsletter, October 23, 2026. wbur.org
• Felix Muzny, Carolyn Jones, Carter Ithier, Hasnain Sikora, Hrutika Harshadbhai Patel and Carla E. Brodley, “A Tool to Map AI Programs in the U.S.: A Snapshot from April 2026 and an Analysis of Requirements for AI Majors and Minors,” arXiv 2606.12428, May 2026. arxiv.org
• Northeastern University Center for Inclusive Computing, “AI Programs in U.S. and Canadian Universities,” live map, read October 30, 2026. cicmap.ai