Who Is Doing the Thinking in This Conversation?

Originally written in Traditional Chinese: 這段對話裡，是誰在想？

Two children can use the same AI and end up as completely different learners. One thinks with the AI. The other hands the thinking over to it. Their homework may look exactly the same.

On October 19, National Taichung University of Education in Taiwan unveiled an "AI textbook" system, and signed letters of intent with the education authorities of Tainan City and Nantou County and an education foundation.

One feature stopped me. The system sorts students' conversations with AI into four types: basic assistance, deep collaboration, basic dependence, and core dependence. Conversations that may show over-reliance on AI are flagged so teachers can step in.

Why? The university's president, Bor-Chen Kuo, pointed to the OECD Digital Education Outlook 2026 and its warning about "cognitive offloading": students hand reasoning, organizing and judgment to AI. The work gets better. The understanding may not. His point, as reported: to reduce students outsourcing their thinking to AI, you first have to be able to see how they use it.

In my course, I teach a principle I call "people before machines." What you need to learn, you do first. What you need done, the machine can do first. Those four types read to me like a ruler for exactly that.

My take: outsourcing thinking does not start when a child opens AI. It starts when they stop thinking first and just ask. We keep asking, "Is my child using AI?" The better question is, "In this conversation, who is doing the thinking?"

To be clear: the news reports did not publish how the four types are judged, so what follows is my own reading, not the system's standard. The partnerships are at the letter-of-intent stage, with schools only beginning to adopt the system and no published outcome data, and I am not endorsing any product. Nor is this about monitoring every message a child sends. What matters to me is the shift from "Should we ban AI?" to "Can we see the thinking?"

One thing to try today. I call it the "who's thinking" check. Open your child's most recent AI conversation, or your own. Before each question, ask: "Did I think about this before asking?" Circle for yes, cross for no. After each answer, ask: "Did I change it, push back on it, or follow up?" Circle or cross again. Two circles: closer to thinking together. Two crosses: the thinking has been handed over.

For parents: tonight, instead of "Did you use AI again?", try "Show me a conversation where you and the AI figured something out together."

For educators: add one line to your next assignment sheet: "What did you ask AI, and what did you change after it answered?"

For professionals: mark yesterday's AI chats with circles and crosses. Are you collaborating, or outsourcing?

Before any system, we can already ask the question ourselves.

Get the Human-First practice card

Four steps, ten minutes, with copy-ready AI prompts. The card and my weekly letter are currently written in Traditional Chinese.

👉 Subscribe free to 碩學苑 (Traditional Chinese)

Follow in English

I also post in English on LinkedIn.

👉 Follow Haishuo Lee on LinkedIn

References