<aside> 📖

In the Decision stage, students move from being observers of the story to becoming active problem solvers and investigators. Rather than using AI simply to provide answers, the system guides students to formulate questions about their chosen animal, gather information, record clues, and gradually understand the relationships among diet, poop characteristics, habitat, and ecological roles.

To accommodate differences in learners’ language abilities, cognitive development, and digital skills, I designed two differentiated AI interaction pathways. Older students interact with the AI through English text-based questioning and organize the information they gather in a Detective Clue Card, while younger students use voice interaction and image uploading to reduce linguistic and operational barriers.

Through this differentiated design, AI functions not simply as an information provider, but as a learning partner that supports observation, questioning, documentation, and decision-making.

</aside>

<aside> 📖

AI as a Learning Partner

</aside>

The upper-primary version introduces an AI agent, “Fly Detective,” as an inquiry partner. Students are encouraged to ask questions in complete English sentences. AI provides information through dialogue and supports them in developing their own questions. This design shifts AI from an “answer machine” to an inquiry partner, while embedding English practice directly into the scientific investigation.

<aside> 📖

From AI Dialogue to Evidence

</aside>

To prevent students from stopping at AI conversation alone, Detective Clue Card transform AI-supported inquiry into structured evidence collection. Students organize the information they gather into categories, turning fragmented AI responses into research evidence that can be used for later analysis and creation. The auto-save function also preserves students’ ongoing work, making the inquiry process itself part of the learning artifact.

<aside> 📖

Multimodal Inquiry for Younger Learners

</aside>

The younger-learner version uses a more accessible voice-based and image-supported interaction. Students can initiate voice conversations with the AI and upload relevant images to their Detective Clue Card as part of their investigation. This reduces the linguistic and operational demands of typing while preserving the core inquiry structure of asking questions → gathering information → recording clues. In this way, learners of different ages can work toward the same learning goals through appropriately differentiated interaction modes.