A study from Technische Universität Berlin’s Science of Intelligence cluster investigated how 90 adult learners responded to automated feedback from a humanoid robot during a learning task involving unfamiliar Swahili instructions.

Key Findings

Researchers compared three feedback approaches: standard feedback after mistakes, feedback adapted to learner performance and enjoyment, and personalized feedback tailored to specific errors. Results revealed important nuances about educational support.

“Robot-delivered feedback helped learners recover from mistakes,” noted the research team, yet “more personalized feedback was not automatically more helpful in the next moment.”

The study found that personalized feedback—while improving overall task performance—sometimes hindered immediate responses due to increased cognitive load. Learners with higher cognitive ability benefited less from task-focused hints, while bored learners responded better to additional guidance.

Implications

The findings suggest effective robot tutoring requires balancing personalization with sensitivity to learners’ moment-to-moment cognitive and emotional states. “The most intelligent tutor may not be the one that always knows what to say, but rather the one that knows when to say less.”

---

Journal: Communications Psychology
DOI: 10.1038/s44271-026-00487-8
Publication Date: June 12, 2026

Source: Technische Universität Berlin, Science of Intelligence cluster