Researchers at Lund University have tested the effect of speech on people's perception of robots.
In previous experiments, scientists have explored how a humanoid robot's eye misbehaviour affects human perception, and it turns out that this leads to a loss of trust in all cases except when the robot made a mistake. The only difference between all these briefcases was what the robot said when he made a mistake.
In order to test the hypothesis, researchers repeated the same experiment they had done in their previous work, but the robot no longer spoke. The pilot was shown a video of a silent and talking robot who had performed acts or made mistakes.
After watching the video, participants were offered a number of questionnaires designed to assess their confidence in the robot, as well as their perception of intelligence, appeal and how much the robot looks alive.
The results showed that, in general, the robots had the greatest confidence, but when the robot could speak, the participants reported that they trusted him almost as well as the robots.
The authors of the study suggest that respondents perceive speech as an indicator of intelligence, and that this is why the level of confidence in such robots is higher, and they plan to test how other factors influence perception, such as the expansion of the pupil and the view.