AI Is the Teaching Assistant Great Teachers Always Needed

The machine can prepare the student. The teacher still has to shape the judgment.

Business schools love turning every new technology into an apocalypse forecast. AI will destroy education. AI will replace professors. AI will make classrooms obsolete. Dramatic. Convenient. Mostly wrong.

The better question is what AI becomes for excellent teachers.

My answer: the teaching assistant we were always supposed to have!

AI, agents, custom GPTs, on demand video, and smarter SaaS are becoming the preparation layer around the course. Not the professor. Not the classroom. The support system. The force multiplier.

I can easily imagine the classic twelve to fifteen session semester staying in place, because time is still time. But around it, an additional twenty to thirty percent of the learning experience becomes AI powered, asynchronous, autonomous, and more personalized. Students review basics before class. They get explanations when they’re stuck. They quiz themselves. They watch short videos. They summarize, compare, rehearse, and arrive with more raw material in their heads.

Good.

So I say, let AI build playlists, explain basics, adapt exercises, summarize content, and help students prepare. Then use class for what class should already be: coaching, mentoring, judgment, friction, pressure, and human guidance.

This is what I’m already testing. Multimedia playlists I create for use outside class. A constantly updated Notion course manual. Targeted agent support doing some heavy lifting. Custom GPTs that help structure milestones and prepare serious individual exams on the school LMS, in lockdown mode.

And this is what I’m seeing: students are relying increasingly on technology for facts. Of course they do. Everyone does. Nobody “challenges” Google Maps before finding a restaurant. You ask where Joe’s place is, then you go.

So when schools tell young students to “challenge AI,” I understand the intention, but I think the phrase is mostly wrong. Bachelor and initial master students are not specialized researchers. They do not naturally confront machines. They use them as answer engines, because that is what the internet, search, apps, and personal computing have trained all of us to do for the last twenty five years.

The issue is not that students have no information. Often, they have too much. They arrive with summaries, definitions, examples, arguments, counterarguments, and whatever else the machine spat out in seven seconds. The problem is that most of it is floating. It has no weight yet. No context. No consequences. No tradeoff. No ownership.

Students often have answers. What they need is the room to hash them out.

That is where class still matters. Not as a delivery mechanism for facts, but as a decision space. A place where students have to explain what they think, defend why it matters, test whether the answer survives contact with reality, and sometimes admit that the shiny AI response was just polished fog.

And that, for me, is the future of education worth building. Not AI replacing teaching. AI extending preparation. Not humans pretending they can outcompute machines. Humans doing what machines still cannot do well: creating the conditions for judgment.

AI gives them information. Teaching helps them decide what it means.