Students need more than answers

In my English classroom at an elementary school in Taitung County, a nod does not always mean understanding. I can ask whether everyone understands, receive reassuring nods and then watch a child wait for classmates to begin before attempting the task. This happens every day.

If nodding were a reliable assessment tool, we could finish the curriculum by Wednesday — just in time for Sports Day.

That gap between appearing to understand and being able to proceed should inform how Taiwan’s schools evaluate artificial intelligence (AI). A clear explanation is valuable, but its educational worth becomes visible in what a student can do.

When I know a student needs support, I pair them with a classmate who understands the task. I become a facilitator, while the helper takes on a similar role. The helper explains, then lets their classmate try, offering guidance along the way.

The work remains the student’s own. The purpose is to help a classmate move through a difficult task, while preserving the opportunity to think through and attempt the task.

This creates two responsibilities. The student must engage with the work. The helper must make an explanation useful to someone else. Knowing an answer becomes the starting point for communication, rather than its conclusion.

My responsibility continues throughout the process. I watch who explains, who attempts the task and where guidance is needed. A correct answer matters, but so does the route by which the pair reach it. If one student supplies the answer and the other merely records it that tells me little about understanding.

There is another distinction to make. Hesitation can signal unfamiliar vocabulary, unclear instructions or uncertainty about how to begin. It does not automatically identify the problem. The teacher must investigate, and remain willing to revise an initial judgement.

These routine decisions are easily overlooked in discussions about educational technology. An explanation can be accurate and still fail to meet a learner’s immediate need. Additional information might help, but sometimes the child needs an opportunity to try with someone nearby.

This leads me to a question for Taiwanese schools considering AI: Does the assistance preserve the student’s opportunity to think, attempt and explain?

A system that supplies a finished response might make an exercise easier to complete. A system used to offer a hint, suggest another explanation or prompt a further attempt could serve a different purpose.

Schools should examine that difference when deciding how a tool should be used in a lesson, if at all. The same principle should shape teacher training. Demonstrating how to generate teaching materials is useful. Teachers also need opportunities to consider what happens after those materials reach children, how misunderstanding becomes visible, when assistance is appropriate and how to check whether a student can proceed.

Peer support is no automatic solution either. Knowing the material does not make a child an effective guide. The helper needs direction, and both students need the teacher’s attention. My classroom offers a perspective on these decisions, rather than a formula for every school.

The wider lesson is that assistance should develop a learner’s capacity to act. In elementary-school classrooms, the value of AI should be judged partly by whether it supports that work. The helper who explains and then allows a classmate to try illustrates something important: Useful knowledge must leave room for the other person’s effort.

AI might know the answer. The educator must still see the person behind the question — and help students learn to see one another.

Casper J.H. Keller teaches English and drama at an elementary school in Taitung County and has more than 25 years of international teaching and leadership experience.