NewsBite analysis · AI, Robotics
The Conversation
When expressive humanoid robots are awkward, people become wary – new brain study
A recent study published in Science Robotics highlights that humanoid robots making errors can lead to increased suspicion among users. Researchers observed participants' brain activity while interacting with the humanoid robot Pepper, which, despite its expressive capabilities, lost users' trust when it made mistakes. The study emphasizes that expressiveness shifts how users perceive robot errors, categorizing them as social violations rather than mere technical faults. This has significant implications for the design and acceptance of robots in everyday environments.
MonitorConfidence: MediumMid term horizon
Published 3 Aug 2026, 12:45 UTC3 min read
Visit original article ↗Lead image — The Conversation
Summary
TL;DR
- Study involved 50 participants interacting with the humanoid robot Pepper
- Errors decreased trust in the robot, regardless of its expressiveness
- Increased oxytocin levels correlated with reduced trust
- Brain activity was monitored using portable sensors during conversations
- Expressive robots' mistakes felt more like social violations
Interesting facts
- Higher oxytocin levels indicated greater suspicion towards expressive robots
- Participants showed increased brain activity in regions tied to uncertainty and social inference
- Study challenges the assumption that expressive robots automatically earn trust