By Wonsuh Song
Today I witnessed a quiet but decisive shift in my social studies content class. For the first time this term, I assigned individual 20‑minute presentations—and one student delivered the entire talk with an AI voice instead of their own. It wasn’t merely a swap of tools; it felt like the ground rules of “what counts as a presentation” were being reset.
The moment sparked a chain reaction. Another student routed their comments and questions through AI as well, and soon we were watching AI present while AI replied—odd to describe, natural to experience. The flow of ideas remained human; only the vocal cords had changed.
One student’s staging stood out. Rather than a single monologue, they built two AI characters and let them converse their way through the content. By trading claims, clarifications, and examples, the “dialogue presentation” lifted attention and comprehension across the room.
It wasn’t flawless. Japanese AI intonation still sounded a bit uncanny. But the willingness to break the convention—“the presenter stands and speaks”—mattered more. Inviting two AI presenters onto the stage for the first time made today genuinely meaningful, fresh, and innovative.
The locus of learning also shifted. In a country‑report assignment, AI shortened the pipeline from research to narration. Students invested less in memorization and recitation, and more in design and verification: refining prompts, planning scene changes, choosing which evidence appears with which voice. That design work became the real site of learning.
Still, twenty straight minutes of AI speech can drag. Today worked because pacing and staging did. A dialogic structure modulated tempo, and every key claim was paired with a brief on‑screen map, table, or quote to re‑capture attention. Next time I’ll ask for a one‑minute human framing at the start and a one‑minute human synthesis at the end, with required evidence reveals in the middle. Submitting prompt logs and draft‑to‑revision trails will make process and authorship assessable.
Equity and ethics must evolve alongside. Students should disclose AI use; fact‑checking and citation should be graded as distinct dimensions. We must caution against voice impersonation and privacy leaks, provide baseline access and guidance, and treat language‑specific intonation limits as a pedagogical parameter. AI can lend efficiency; humans still stage meaning.
Today I saw the first real step in how university teaching and student presentations might evolve. I intend to keep encouraging bold attempts and widening the space for experiment, missteps included. Classroom rules are not fixed; they change on the day someone tries something new—and makes it work.
Wonsuh Song (Ph.D.)
Full-time Lecturer, Shumei University / NKNGO Forum Representative
댓글 남기기