Why Educators Must Reclaim the AI Conversation

A reflection on lessons learned building at the intersection of AI and education

When I started building Bitta.ai, I thought I was solving a technology problem. A few months in, I realized I was solving a trust problem.

Here is what I kept seeing: students were dragging and dropping their educators’ lecture slides into ChatGPT and Claude, getting summaries, and moving on. On the surface, that looks like efficiency. But something important was getting lost in the middle — the educator’s intent, context, and voice. And perhaps more critically, the feedback loop that should flow back to the educator was flowing instead to OpenAI and Anthropic.

That realization changed how I think about AI in education entirely.

The Problem With Unguarded AI

When a student asks ChatGPT a question about course material, a few things happen simultaneously. The AI answers — sometimes correctly, sometimes not. The student accepts the answer — sometimes critically, sometimes not. And the educator learns nothing.

There is no record of what confused the student. No signal about which concept needed more reinforcement. No visibility into the gaps between what was taught and what was understood.

The feedback loop — one of the most powerful tools an educator has — simply disappears.

What Control Actually Means

Control is not about restriction. It is about intentionality.

When an educator builds their own AI representative — trained on their own materials, shaped by their own methodology, bounded by their own guardrails — something fundamentally different happens. The AI speaks in their voice. It reflects their priorities. It knows what to emphasize and what to deflect.

And crucially — every question a student asks, every concept they struggle with, every moment of confusion — gets recorded, analyzed, and returned to the educator. Not to a tech company. To the person who is actually responsible for the student’s learning.

That is what a guarded environment makes possible.

The Practice Loop

One of the most underestimated applications I have seen is mock practice. Sales roleplay. Negotiation scenarios. Interview simulations. These are experiences that traditionally require a partner, a coach, or a scheduled session.

With an AI representative built by an educator who actually understands the subject — an educator who has coached hundreds of students through real negotiations, a career director who hired at Meta for eight years — the practice becomes genuinely useful. Students can repeat it at midnight before an exam. They can fail safely and try again immediately.

And the educator can review every transcript. Not to surveil — but to learn. To see which objections students struggle with. Which scenarios expose real gaps. Which parts of the coaching need refinement.

The Lesson I Keep Coming Back To

AI will be in every classroom whether educators invite it or not. The question is whether educators are the ones shaping the experience — or watching from the outside while someone else does.

The tools exist. The insight is available. The only thing missing is intentionality. That is worth protecting.

Siyuan Guo is the founder of Bitta,ai, an AI-powered platform and presentation tool that transforms static content, like PDFs, slides, and websites, into interactive AI agents that can present your ideas, answer questions, and engage with audiences dynamically. This piece was originally published on the site of our Worthy partner, Mentor.net. and is shared here with permission. You can reach out to Siyuan via email here.