Recently, Italian researchers conducted a study in which they gave robots the ability to "think out loud" to help human users better understand the decision-making processes of robots.
The main member of the study, Arianna Pipitone, said: "Human inner speech is connected to the subconscious, so we wanted to investigate this connection in robots."
Researchers programmed the Pepper robot to verbalize its thought processes. This means that Pepper is no longer a "black box," and its underlying decisions are more transparent to users.
In the experiment, researchers tried to explore how this inner speech affects robot behavior. In one instance, Pepper helps people set the table according to etiquette. When the user asks Pepper to violate etiquette by placing the napkin in the wrong place, Pepper starts to talk to itself and wonders whether it should continue with the action.
Once the user confirms his request, Pepper says to itself: "This situation makes me uncomfortable. I shouldn't break the rules, but I can't upset him, so I'm doing what he wants," and places the napkin where requested.
Allowing robots to "think out loud" is helpful when they ignore requests because the robot can articulate the reasons for not executing commands. For example, a specific object is unreachable, the requested motion is not feasible, or a component is not functioning properly.
According to the study published in the journal iScience, by comparing Pepper's performance with and without the use of internal language, researchers found that Pepper has a higher task completion rate when engaging in self-dialogue.
Moreover, this internal voice capability could be even more useful in collaborative situations between robots and humans, such as care robots.
Reference：
https://www.theguardian.com/technology/2021/apr/21/study-inner-life-ai-robot-thinks-out-loud?fbclid=IwAR3sZflAMCl3SODsj42y5DLS-lc9xt3gV9AtY8FE5CExQEAiuqIsEv3qVT8
.jpg)
