will AI replace self-enrichment teachers?

Twenty-four of your thirty core tasks show zero AI penetration, according to O*NET task data. That's not a small margin. Tasks like conducting live classes, supervising field trips, meeting with other instructors about individual students, and guiding people through hands-on experiential learning all require you to be physically present, read the room, and respond in real time. No tool does that.

The relationship side of this work is where your irreplaceability really shows. Your students aren't just trying to absorb information. They're trying to learn a skill that matters to them personally, whether that's sourdough baking, oil painting, or personal finance. They need encouragement when they mess up the first attempt. They need you to notice that someone at the back of the room looks confused but won't raise their hand. That kind of moment-to-moment human attunement is not something any current AI system can deliver in a live setting.

Enforcing class norms, managing group dynamics, planning guest speakers, and grading work with personal feedback all stay in your hands. And the motivational element of self-enrichment teaching is significant. People sign up for your class voluntarily, often after a long work week. Keeping them engaged, returning week after week, and actually improving at something they care about depends almost entirely on you. That's your strongest card.

AI is genuinely bad at anything that requires physical presence and live adaptation. In self-enrichment teaching, your classes often involve demonstration, hands-on practice, and real-time correction. If a student's knife grip is wrong in a cooking class, or their posture is off in a dance class, an AI system can't see that and respond to it the way you can standing two feet away.

AI also struggles with the emotional texture of adult learning. Adults in enrichment classes often carry anxiety about looking foolish in front of others. They have reasons for being there that go beyond the skill itself, a recently retired person taking up watercolour, someone using a writing class to process a life change. Reading that subtext and adjusting your approach is not something AI can do. It doesn't know what's unspoken, and it has no way to find out.

There's also an accountability gap. If an AI-drafted lesson plan is poorly calibrated for your group's actual skill level, you're the one managing the fallout in real time. AI can't be responsible for a class that went badly. You can. That distinction matters in a profession built on trust and personal relationship.

The tasks where AI gets traction are the ones that happen before or after you're in front of your students. Writing instructional articles, drafting lesson plan outlines, preparing program objectives: these are real time sinks that AI tools handle competently. Tools like ChatGPT and Claude can produce a structured 6-week pottery course outline in a few minutes, something that would have taken an afternoon to draft from scratch.

For multimedia preparation, tools like Canva's AI features and Adobe Firefly can generate visual aids, slide decks, and course materials faster than building them manually. If you teach cooking, photography, or fitness, you can use these to put together handouts and reference sheets without needing a graphic design background. That's genuinely useful.

On the curriculum review side, AI can help you analyse student feedback at scale. If you collect written evaluations from 30 students, tools like Notion AI or even a simple ChatGPT prompt can pull out recurring themes and flag weak spots in a course much faster than reading each one individually. The Anthropic Economic Index rates tasks like writing instructional content and preparing lesson outlines at above 85% AI penetration for roles like yours, which means these are the tasks to hand off first. But note what's missing from that list: everything that happens in the actual room with actual people.

The biggest shift is in prep time. What used to take a Sunday afternoon, building a new unit outline, writing up session notes, drafting a handout, now takes maybe 45 minutes with AI doing the first draft. You're spending less time staring at a blank page and more time refining and making things yours.

What hasn't changed at all is the teaching itself. You're still arriving early to set up, reading the energy in the room from the first five minutes, adjusting your plan when three people missed last week's session, and staying late to talk with the student who's struggling. That rhythm is identical to five years ago. The pre-class and post-class admin is lighter. The class itself is not.

You're also probably spending more time on the parts that matter most, the actual instruction and the personal interactions, because the admin overhead has dropped. If you're running your own classes independently, the marketing side has also changed: AI tools can help you write class descriptions, social posts, and email newsletters faster. That frees up time you'd have spent on copy to spend on curriculum.

The BLS projects 3.7% growth for self-enrichment teachers through 2034, which translates to roughly 51,400 job openings per year across a base of 417,500 employed workers. That's a steady market, not a shrinking one. And the openings number is high relative to the base, which means there's real turnover and entry opportunity, not just growth at the margins.

The demand here is structural. People want to learn skills in person, with guidance, in a social setting. Online video tutorials have existed for over a decade and they haven't replaced enrichment classes. AI-generated instructional content won't either, for the same reasons. The appeal of a cooking class or a watercolour workshop isn't just the information. It's the experience of doing it with other people under someone's guidance. That's durable demand.

AI's low exposure score of 9% for this role means the growth isn't being eaten by automation. The new jobs are real jobs, not roles that will quietly shrink once AI gets better at the margins. If anything, AI handling the prep work makes it slightly easier to run more classes or take on more students, which could support earnings growth for independent instructors without increasing their hours.

The 9% exposure score is unlikely to rise much. The tasks AI handles well in this role are already being handled. Writing lesson plans and instructional content is the ceiling, not a stepping stone to automating classroom delivery. For AI to genuinely threaten the live instruction part of your job, it would need to replicate physical presence, real-time human feedback, and social group dynamics. That's not a near-term problem. Five years from now, the prep tools will be better, but the job itself will look essentially the same.

The one area to watch is AI-powered video instruction. Products like Synthesia can generate video lessons from a script, and some enrichment platforms are experimenting with AI-guided self-paced modules. But these compete with YouTube tutorials, not with live classes. The people who choose in-person enrichment are choosing it deliberately. As long as that preference exists, and the BLS data suggests it does, your position is stable. Ten years out, the exposure score might inch up to 15-20% as AI gets better at adaptive content. That's still a role with 80% human-owned tasks.

The clearest thing to double down on is the live instruction itself. The 24 zero-penetration tasks are your professional core. Classes that involve physical skill development, real-time group interaction, experiential activities, and guest speakers are exactly where you want to build depth and reputation. The more your teaching is defined by what happens in the room, the safer your position is.

If you're running independent classes, get comfortable using AI for the prep side but keep your voice in the final output. A lesson plan drafted by AI and then shaped by three years of knowing your students is better than either alone. The same goes for marketing copy, student newsletters, and course descriptions. Let AI do the first draft. You finish it.

On the skills side, the instructors who'll do best over the next decade are those who are genuinely excellent at teaching adults, not just knowledgeable about their subject. Adult learning theory, group facilitation, and motivational techniques are skills worth developing deliberately. Look at courses from the Association for Talent Development or the Learning and Performance Institute if you want structured training. These aren't things AI will make irrelevant. They're things AI will make more visible, because the instructors who have them will stand out clearly from those who don't.

Finally, consider the breadth of what you teach. The self-enrichment field covers cooking, fitness, photography, writing, finance, dance, and more. Teachers who can connect skills across domains, or who bring a distinctive personal perspective to a subject, build audiences that stick. That's hard to replicate and harder to automate.