Robot embodiment changes verification and debugging: learners treat robot execution as perceptual evidence for hypothesis-driven revisions, while AI-Only learners rely on generic help-seeking¶

Claim · All claims
Evidence · 1 study (2 entries) · causal r2 · q2
Currency · newest study 2026 · checked against its sources 2026-10-20 (automated, 2 of 2 entries) · not reviewed by a person

Subclaims¶

q2 i? In the AI+Robot condition, participants initiated robot execution early and used the robot's motion as a concrete trace to localize failures and ask the AI about specific steps. → Leyi Li 2026 q2 i? In the AI-Only condition, help-seeking sometimes became generic and repetitive, and several students progressed with partially incorrect reasoning without a perceptual signal prompting revision. → Leyi Li 2026 (2)

Evidence¶

Leyi Li 2026¶

Leyi Li, Chenyu Du, Jiafei Sun, Erick Purwanto, and Qing Zhang. (2026). RoboBlockly Studio: Conversational Block Programming with Embodied Robot Feedback for Computational Thinking. Designing Interactive Systems Conference (DIS '26). https://doi.org/10.1145/3800645.3813071

q2 · i? · design · r2

Qualitative observation and interviews in the counterbalanced study: in the AI+Robot condition, "students often watched the run to localize where a failure emerged" and then asked the AI about a specific step or block choice, followed by verbalized reasoning.

"Rather than only asking whether a solution was "correct, " students often watched the run to localize where a failure emerged (e.g., an incorrect turn, a collision with an obstacle, or a missed target region), and then revised the corresponding block segment."

Leyi Li 2026 (2)¶

Leyi Li, Chenyu Du, Jiafei Sun, Erick Purwanto, and Qing Zhang. (2026). RoboBlockly Studio: Conversational Block Programming with Embodied Robot Feedback for Computational Thinking. Designing Interactive Systems Conference (DIS '26). https://doi.org/10.1145/3800645.3813071

q2 · i? · causal · r2

In the AI-Only condition, learners verified programs through on-screen inspection and AI dialogue; the authors observed that "several students in the AI–Only condition progressed further with partially incorrect reasoning" without a salient failure signal.

"we observed that several students in the AI–Only condition progressed further with partially incorrect reasoning, because there was no immediate, perceptually salient signal (e.g., a visible misstep or failed interaction) to prompt targeted revision."

Discussion¶

Related Claims¶

• 30 of 32 participants preferred the AI+Robot condition, describing the robot as making success and failure visible, though a minority found AI-Only useful for focusing on block construction — related
• Embodied collaborative construction can build shared geometric reasoning — a broader claim this one bears on
• Within a single LLM session, students' help-seeking shifts toward answer- and verification-seeking while explicit self-regulation modestly increases — related
• Students' help requests to the AI tutor were predominantly instrumental rather than executive, with procedural content and low verification — related
• Scaffolded task progression shifts novice reasoning from local completion toward diagnosis and systematic debugging as task demands increase — related
• A language-based AI facilitator has access only to verbal traces of embodied inquiry and steers inquiry toward what can be put into words — related