For decades, community colleges have helped turn students’ ambition into employment by prioritizing practical skills over abstract theory. Most students enroll in two-year colleges because they want to build a career and provide for their families, and instructors design programs and curricula to meet that goal. In practice, that has always meant hands-on learning: giving students real opportunities to apply what they know rather than leaving them stranded in the theoretical.
But while designing genuinely hands-on learning required plenty of care even when a lot of work was done by hand, it’s only gotten harder in an era where community colleges are often preparing students for work in the knowledge economy. To put it another way, in a world increasingly shaped by AI, designing effective hands-on learning is becoming both harder and more essential. The tools students will use on the job are changing faster than most curricula can be rewritten, which forces every instructor to ask what "hands-on" should even mean now.
Consider cybersecurity, a field I have taught in for nearly a decade and one where AI is reshaping day-to-day work most dramatically. Three years ago, hands-on cybersecurity training would have meant asking a student to configure a security tool, emulate some behavior such as running a malicious file, then looking at how that activity is detected by the tool. Today, though, AI would be present at every step of the process: helping configure the tool, running the malicious behavior, and evaluating the tool’s performance. That means some of the hands-on parts of the learning experience would be taken over by AI. It also means students would be responsible earlier on for a different, but equally important, task: using their own judgment and critical thinking to evaluate how the AI reached the outputs it did, and whether those outputs are correct, secure, and fit for purpose.
Examples like this demonstrate what many instructors have already seen: in the AI era, hands-on learning is becoming as much about critical thinking as it is about technical execution. In many ways, this shift is an exciting one. After all, isn’t critical thinking one of the most important things a student should get from their higher education experience? But for busy instructors reimagining their curricula in real time, there’s also a real risk: that we mistake using AI for understanding core concepts, and graduate students who can complete assignments, but cannot apply independent analytical thinking. After all, it’s easy to give students tasks and assignments that ask them to produce code or other materials using AI. That might feel like “hands-on” learning for the students doing it, but if we aren’t imparting enough judgment and critical thinking skills, students won’t leave with the ability to actually put AI into practice.
The mandate for community colleges is the opposite: to ensure students master core principles first, and only then teach them to apply AI to work faster and think bigger. How does that square with hands-on, experiential learning?
Take a common task in our field: writing a Python script to integrate a security tool. A student who "vibe codes" that script without grasping fundamental programming concepts cannot troubleshoot it when it breaks, or be confident it is doing what it is supposed to do in the way intended. A student who learns the core concepts first can use AI to move faster during development, catch the model's mistakes, and even carry those fundamentals with them when they learn a new language. That student leaves not just more employable, but a sharper critical thinker — exactly what employers in an AI-enabled economy are looking for.
So how should institutions build this kind of learning?
First, work backwards from action-oriented objectives. Decide what students should be able to do on the job, then design the course to get them there, rather than starting from a textbook and hoping it translates. My course, for instance, is built on the curriculum and approach developed by the nonprofit CodePath, which has brought in industry experts to build a cybersecurity program that feels up-to-the-minute with the skills that students will actually need to secure (and succeed in) jobs in the cybersecurity industry.
Then, create relevant tasks and scenarios. Build realistic exercises that map directly to those objectives, so students practice the judgment they will actually need — including when to use an AI tool and when to avoid it. CodePath uses a system for this made up of green flags (e.g., use AI to clarify unfamiliar terms, explain concepts, or check for errors) and red flags (don’t trust AI blindly or replace your own decision-making) that my students can rely on throughout the course.
Finally, focus lectures on the "why," not the "how." Teach the durable concepts that let students adapt as tools change, rather than step-by-step instructions that expire the moment the software updates.
None of this means slowing down to resist AI; instead, it means being deliberate about the order in which students learn. Community colleges have always been where practical skills meet real opportunity, and getting this balance right will determine whether our students merely keep up with AI or learn to lead with it. The definition of hands-on is changing, but the mission behind it has not: to send students into the workforce ready to do the job, and ready to think.
